1             ABSTRACT

Despite being a very common term in Neuroscience research, a lot of ambiguity persists in the literature regarding the precise definition of top-down control. In this review, we propose a more rigorous model of ‘top-down control’ as the integration of information contingent upon the maturation of neuronal ensembles. This model is explored in negative and positive valence studies that have investigated the medial prefrontal cortex (mPFC), an important heteromodal association cortex that is related to goal-directed behavior. In face of the new definition, we conclude that the maturation of neuronal ensemble in the mPFC is necessary for goal-directed behavior. We posit that a focus on the mechanisms of ensemble maturation could become a unifying facet of future research around the mPFC, allowing different lines of neuroscientific investigation to contribute to one another.

2             TOP-DOWN CONTROL IN NEUROSCIENCE RESEARCH

The definition of ‘top-down’ and ‘bottom-up’ models has been widely adopted by many scientific fields, with different and often contradictory meanings amongst them. In the field of Neuroscience and Psychology, the term ‘top-down’ is commonly used as jargon in scientific papers but rarely actually defined (Rauss & Pourtois, 2013), which likely stems from a lack of consensus on a rigorous definition for top-down processing. As a result, the term is often used in contradictory ways. For instance, ‘top-down control’ has been used as a defining characteristic of the visual processing in V1 in anesthetized ferrets (Roland et al., 2006), of the stress-regulating influence of the mPFC over the thalamus-BNST-amygdala pathway in rodents (de Kloet, de Kloet, de Kloet, & de Kloet, 2019) and for the role the parietal cortex in attention orienting in primates (Shomstein, 2012) – the same term used for different species, different states of consciousness, different brain regions entirely. It could be argued that such a definition would be rendered useless due to its broadness in scope.

To have a more precise definition of top-down control in the context of Neuroscience, it is necessary to think in terms of hierarchies in information encoding. One of the fundamental functions of the nervous system is to perform information processing, taking complex environmental and interoceptive inputs and allowing the organism to perform actions in accordance with its environment – a process coined as the ‘perception-action cycle’ (Fuster, 2001). As the information flows from the peripheral nervous system to the spine to sensory cortices to association cortices, information is encoded via mechanisms that are intra-neuronal (e.g. changes in gene expression, receptor expression, and spine morphology) and extra-neuronal (e.g. myelin plasticity) (Tozzi, 2015). Changes in neuronal activity promote changes in neuronal connectivity, resulting in the formation of ‘neuronal ensembles’ or ‘memory traces’– biological substrates that encode a particular memory (Thompson, 2005).

We propose a definition of top-down and bottom-up processing as follows: at lower levels of the hierarchy (e.g. peripheral nervous system or PNS) bottom-up processing occurs and the information is processed at a greater level of detail. On the other hand, at higher levels of the hierarchy (e.g. association cortices), top-down processing takes place, meaning that the incoming information is integrated (Figure 1). What differentiates hierarchical levels is their relative sparse connectivity: the information encoded by many ensembles in lower hierarchical levels is condensed into fewer ensembles in higher hierarchical levels. The condensation of information is the characteristic that allows information to be integrated from multiple inputs in higher hierarchical levels. Importantly, this model does not pose that only frontal cortices exert top-down control (e.g. definition adopted by White et al., 2018). For example, sensory cortices may exert top-down control over afferent spinal inputs. Moreover, the same brain region can have top-down and bottom-up processing occurring simultaneously: for instance, a sensory cortex can exert top-down control over the spinal inputs while providing efferent bottom-up signals to an association cortex.

Critically, this definition differs from the general way it is used in literature: some authors have asserted that the activity of neocortical regions implies top-down control over subcortical structures (Chiesa, Serretti, & Jakobsen, 2012). Under our proposed definition, the mere simultaneous firing of a brain region at a higher hierarchical level with another region at a lower hierarchical level does not necessarily imply top-down control. Rather, top-down control occurs at an ensemble level, contingent upon changes in cellular activity and connectivity. Since brain regions at higher hierarchical levels (e.g. heteromodal association cortices) have sparse connections with many other brain regions, we propose that their capacity to integrate information does not occur immediately. Instead, the memory ensemble undergoes a process of maturation, in which the connections in higher hierarchical levels are gradually strengthened over a period of time (see Section 4.3).

3             GOAL-DIRECTED BEHAVIOR AND THE MPFC

As mammals evolved, their actions became more complex – i.e. based less on simple stimulus-response loops and contingent on prior experience (Carlén, 2017). The ability of an organism to appropriately modify its actions to optimize the possible outcomes in a given scenario has been coined as goal-directed behavior (Zwosta et al., 2015). Goal-directed behavior is uniquely different from innate reflexes or habitual actions because there is no predetermined set of actions which could be constructed ex-ante. Instead, the organism needs to promptly adapt its actions based on constantly changing environmental stimuli. (Verschure, Pennartz, & Pezzulo, 2014)

Goal-direction involves brain-wide networks and therefore no single brain region should be considered a ‘goal-direction center’ of the brain. However, the capacity to integrate multimodal forms of input is paramount for animals to behave sensibly to changes in their environment. Therefore, goal-direction is contingent upon the exertion of top-down control from associative cortices over sensory and limbic cortices (see Section 2). While other associative cortices, such as the parietal cortex (Cohen, 2009), have been related to goal-direction, this review will mainly focus on the medial prefrontal cortex (mPFC).

The mPFC can be subdivided into the dorsomedial prefrontal cortex (dmPFC), which constitutes the anterior cingulate (AC) and the most dorsal section of the prelimbic cortex; and the ventromedial prefrontal cortex (vmPFC), which constitutes the infralimbic cortex (IL) and the ventral-most section of the prelimbic cortex (Figure 2) (Uylings & Van Eden, 1991). The vmPFC receives more limbic projections and processes emotional and interoceptive information while the dmPFC has more connections with sensory and motor regions (Heidbreder & Groenewegen, 2003).

The mPFC integrates motor information, exogenous stimuli (incoming mainly from thalamus), and endogenous stimuli (incoming from connections with the limbic system, which includes the amygdala, hippocampus, and nucleus accumbens) (Kamigaki, 2019). Therefore, the mPFC is anatomically positioned to integrate multiple modes of information and to modulate behavior (Gazzaley & Nobre, 2012). This rich anatomical connectivity allows the mPFC to act as an important hub for goal-directed behaviors, allowing the association between certain actions to positive outcomes and others to negative outcomes, thereby increasing the organism’s adaptability over time. (Kamigaki, 2019)

In behavioral neuroscience, two broad types of paradigms can be used: Pavlovian or instrumental. In pavlovian conditioning, the animal learns an association between two stimuli (e.g. a sound and a food reward). In contrast, in instrumental conditioning, the animal associates a self-initiated behavior with a stimulus (e.g. a nose poke with a food reward). Furthermore, paradigms can also be defined by the valence of their stimuli: a stimulus can be appetitive if the outcome is a reward (e.g. a food reward) or aversive if the outcome is a punishment (e.g. a foot shock). The valence of the outcome is important because it primes the attention of the organism towards the context: it is often the case that a neutral outcome does not form a robust memory (Lonsdorf et al., 2017). Therefore, in both Pavlovian and instrumental setups, the valence of a stimulus (whether positive or negative) has an impact on the memory formed: in Pavlovian conditioning, the valence potentiates the association between both stimuli, while in instrumental conditioning, it modifies the likelihood of the animal performing the same behavior in the future.

The following sections are a discussion of two subtypes of study which have investigated the mPFC: Pavlovian-negative (focusing mainly on fear-conditioning paradigms) and Instrumental-positive (focusing on addiction studies and the delay-discounting task). It is important to note that they are extremes in terms of training complexity: fear conditioning might require only one session to establish a memory that will last for the entire lifetime of the animal (Gale et al., 2004), whereas training in the delay-discounting task might take several weeks (Robbins, 2007). Despite this striking difference, we will propose that ensemble maturation is a unifying characteristic of both types of study. We will describe the basic circuitry involved in positive and negative valence studies, followed by a delineation of the relationship between mPFC, top-down control, and goal-direction in each study type.

4             TOP-DOWN CONTROL IN NEGATIVE VALENCE STUDIES

4.1        INTRODUCTION TO FEAR CONDITIONING

The expression of fear offers evolutionary advantages for animals and can be construed as an aspect of goal-direction: the brain must associate environmental cues with negative valence stimuli, which allows the animal to adapt its behavior in a future encounter with the same environment in order to optimize possible outcomes. Moreover, this process of association has an element of uncertainty because no organism can encounter every possible environment. Instead, animals need to have a model of the world which is based on previous experiential evidence (Rusu & Pennartz, 2019). The organism not only learns which environments are safe or unsafe, but it also uses this information to inform the behavioral decisions upon encountering new and unknown environments. Furthermore, animals also need to be flexible and be able to extinct fear memories, because environments which were once threatening in the past may be safe in the future (Moscarello & Maren, 2018).

A paradigm devised to model this natural phenomenon and to explore the mechanisms of associative learning is fear conditioning. Fear conditioning involves the association of a neutral conditioned stimulus (CS) paired with an aversive unconditioned stimulus (US). The animal is placed in an operant box for the first time and it receives a foot shock a few minutes later – the pairing of CS and US is called fear acquisition. In auditory fear conditioning, the CS is a tone and in contextual fear conditioning, the CS is the contextual information. After fear acquisition, the animal is subsequently provided with the same CS, but this time it does not receive a foot shock – this process is named extinction training. Importantly, during extinction, the original fear memory is not erased, but rather a new competing memory is established (An et al., 2017), which explains why extinction training suppresses the fear memory only transiently (Bouton, 2004) and in a context-dependent manner (Bouton & Bolles, 1979). Posterior to extinction training, the animal is confronted with the original CS again to test if the fear memory is reinstated after the extinction phase – which is a process known as ‘renewal’. In all stages of fear conditioning, the animal’s freezing behavior is used as a proxy of the underlying fear memory.

The mPFC seems to be important in two moments of fear conditioning: renewal, which has been mostly associated with the PL, and extinction, which has mostly implicated the IL (Knapska & Maren, 2009; Stern, Gazarini, Vanvossen, Hames, & Bertoglio, 2014). For decades, with increasingly advancing methods, the goal of this field of research has been to unveil the underlying ‘fear circuitry’, which has been found to involve the interaction between the mPFC, amygdala, and hippocampus. (Maren, Phan, & Liberzon, 2013)

4.2        THE INTERPLAY BETWEEN AMYGDALA AND MPFC

The amygdala can be subdivided into two main nuclei: the basolateral amygdala (BLA), which constitutes the lateral, basal and basomedial nuclei; and the central nucleus (CeA), which constitutes lateral and medial subregions (Pitkänen, Savander, & LeDoux, 1997). The BLA receives sensory inputs via thalamus and it receives projections from neocortical structures, such as the hippocampus and the mPFC (Pitkänen, Pikkarainen, Nurminen, & Ylinen, 2006). The BLA has long been thought of as the site in which the pairing of CS and US would occur (Marek, Sun, & Sah, 2019), while the CeA has been considered the main output of conditioned fear responses. Connecting the BLA and CeA is a group of GABAergic neurons called intercalated cells (ITC) (Duvarci & Pare, 2014).

One of the first examples of the interaction between the mPFC and the amygdala in a fear conditioning paradigm was the observation that a lesion in the vmPFC induced impairments in extinction (Morgan, Romanski, & LeDoux, 1993). It has been later shown that in the course of extinction learning, there is a reduction in synaptic efficacy in glutamatergic neurons of the mPFC that project to principal neurons in the BLA; however, the synaptic efficacy of mPFC projections to ITCs remains unchanged, leading to enhanced inhibition of the central nucleus of the amygdala (Cho, Deisseroth, & Bolshakov, 2013). It has also been shown that extinction can be facilitated by the induction of synaptic depression of a monosynaptic projection between BLA and mPFC (Klavir, Prigge, Sarel, Paz, & Yizhar, 2017). Furthermore, optogenetically silencing projections from the IL to the BLA impairs extinction learning while optogenetically stimulation of this pathway enhances extinction learning. (Bukalo et al., 2015)

Although there is a bidirectional anatomical connection between mPFC and amygdala, this does not entail that the mPFC always exerts top-down control over the amygdala. For instance, the behavioral response of freezing itself, immediately after acquisition, is mainly driven by the amygdala and periaqueductal gray function (Herry & Johansen, 2014). The disruption of mPFC activity in the initial fear acquisition stage does not inhibit the freezing response (Gilmartin, Balderston, & Helmstetter, 2014; Heroux, Robinson-Drummer, Sanders, Rosen, & Stanton, 2017; Lee & Choi, 2012; Zelinski, Hong, Tyndall, Halsall, & McDonald, 2010), suggesting that the initial fear expression is not mediated by mPFC function.

However, the function of goal-direction of the mPFC in fear conditioning lays upon the fact that the expression of fear is not appropriate under all circumstances. Upon encountering a possibly threatening environment, the organism needs to recall the context in which it previously had negative experiences and then prime an adaptive response – i.e. to perceive the new context as high-threat and flee/freeze or to perceive it as low-threat and behave in a normal, exploratory way. The choice of behavior from a complex repertoire must be finely controlled and is not only based on current stimuli, but also on the possible imminent threats that could occur the next moment (Giustino & Maren, 2015). Therefore, the mPFC could be considered as a goal-direction hub in fear conditioning, not because of the expression of fear itself, but rather determining when to express fear as the most appropriate behavior. (Moscarello & Maren, 2018)

4.3        THE INTERPLAY BETWEEN HIPPOCAMPUS AND MPFC

The hippocampus is thought to provide contextual information, as evidenced by the fact that hippocampal lesions lead to impaired fear expression when a foot shock is paired with a context, but not when it is paired to an auditory tone (Phillips & LeDoux, 1992). More recent research has corroborated that idea, showing that hippocampal neurons are preferentially active during context presentation, independent of whether the animal is immediately shocked or previously fear conditioned. (Zelikowsky, Hersman, Chawla, Barnes, & Fanselow, 2014). Although the relationship between hippocampus and mPFC has long been thought as one of excitatory feed-forward excitation (Padilla-Coreano et al., 2016), recent evidence has suggested that more important for fear extinction is the feedforward inhibition of the hippocampus to PV-positive interneurons in the IL, but not somatostatin-positive interneurons or principal glutamatergic neurons in the IL (Marek et al., 2018).

Outside the field of fear conditioning studies, the interplay between hippocampus and mPFC has been traditionally been thought as one of the supporters of memory consolidation (Alvarez & Squire, 1994): the hippocampus and its surrounding entorhinal cortex would function as the neural substrates for recent memories while the mPFC and other neocortical regions would serve as substrates for remote memories (McClelland, 2013), although the exact mechanisms of systems consolidation remained elusive for a long time.

In a recent seminal paper, Kitamura et al. used ensemble-specific techniques to demonstrate a possible mechanism to the consolidation of fear memory in the mPFC over time. Rather than being initially encoded in the hippocampus and later transitioning to mPFC, they found that there were mPFC ensembles formed during the initial stage of fear acquisition (Kitamura et al., 2017). These mPFC ensembles were found to receive projections from both the hippocampus and the BLA, indicating that they could be sites of top-down control over the incoming bottom-up signals. These mPFC ensembles, however, were initially immature and were not naturally activated by environmental cues, but they could be artificially activated with optogenetics to induce freezing (Kitamura et al., 2017).  

The authors also found an opposing effect of maturation between mPFC and hippocampus: they found that the initial immature mPFC ensemble became mature within 14 days, and the initially active hippocampal ensemble was no longer involved after 14 days. Meanwhile, BLA ensembles were persistent throughout the entire period, which suggests its role in encoding the valence of a fear memory (Kitamura et al., 2017). An interesting hypothesis to explain the transition between a hippocampal-driven memory and a neocortical-driven memory is that the hippocampus is constantly creating new neurons, which could integrate into the hippocampal network and disrupt established memory ensembles. (Kitamura & Inokuchi, 2014)

This study helps to answer an important question discussed previously (see Section 4.2): why is mPFC activity not important for the initial expression of fear, but rather the initial fear expression seems to be driven mainly by subcortical structures? One possible answer that the association of inputs from lower hierarchical levels (hippocampus and amygdala) is not initially strong enough to allow mPFC ensembles to exert top-down control over the primed bottom-up signals. These mPFC ensembles need time to become mature (Figure 3) and in the context of fear conditioning, this maturation takes around 14 days and results on increased complexity in dendritic morphology in these ensemble cells (Kitamura et al., 2017).

4.4        LIMITATIONS OF FEAR CONDITIONING

An important limitation of the fear conditioning paradigm is its dependence on fear and pain circuits (Herry & Johansen, 2014), which results in memory effects that are not generalizable to other forms of learning. This can be exemplified with the phenomenon of incubation, in which the fear response can be potentiated over time without extra training, which is a phenomenon known as incubation (Eysenck, 1968).

Furthermore, stress itself has been shown as a confounding factor in fear conditioning: exposure to chronic stress enhances fear expression (Maroun et al., 2013) and it is correlated with a decrease in dendritic morphological complexity in the mPFC. (Izquierdo, Wellman, & Holmes, 2006). A recent study has demonstrated that mild-fear conditioning promotes the formation of an ensemble in the mPFC, whose activation is sufficient to induce fear expression one month later. However, a strong fear conditioning (using three foot shocks, instead of one) does not induce the formation of mPFC ensembles (Matos et al., 2019).

5             TOP-DOWN CONTROL IN POSITIVE VALENCE STUDIES

In the field of positive valence paradigms, it has been consistently found that mPFC impairments do not necessarily decrease the performance of the animal in the tasks. However, mPFC impairments usually have a negative impact on aspects related to task switching and cognitive flexibility (Floresco, Block, & Tse, 2008). This general phenomenon relates to the importance of the mPFC in the integration of heteromodal information to allow goal-directed behavior.

Similar to negative valence studies, the term top-down control is often used in positive valence studies to describe the role of the mPFC in attention and impulsivity (Miller & D’Esposito, 2005). We posit that this general definition is not correct: instead, similarly to the presented evidence from aversive stimuli studies (see Section 4.3), top-down control should be mediated by mechanisms of ensemble maturation in the mPFC.

5.1        ADDICTION STUDIES

Addiction studies are an all-encompassing definition for paradigms that use addictive substances as positive valence stimuli. One example is the forced-abstinence paradigm, in which the animal goes through a period of training in which it associates an action (e.g. a lever press) with a drug reward. After this initial training period, the drug reward is paired with a negative stimulus (e.g. a foot shock). The goal of this type of study is to assess the control of impulsivity, observing if the animal is capable of refraining from a short-term reward (drug reward) because of its longer-term consequences (after a while, it will receive a foot shock as a punishment).

Using this type of paradigm, it has been shown that long-term cocaine self-administration reduces PL excitability, which can be rescued with optogenetic stimulation. (Chen et al., 2013) This suggests that hypoactivity of mPFC is related to a loss of top-down control and consequent compulsive drug seeking. Furthermore, a recent study has shown that when pairing a lever press to a punishment, there is a notable shift in synaptic plasticity in the vmPFC neurons which project to the shell of the nucleus accumbens (NAc) (Halladay et al., 2020).

While the amygdala has a very clear role in encoding ‘fear memories’ in the context of negative valence paradigms, it has also been implicated in the processing of positive valence stimuli. An increase in neuronal firing of the BLA is necessary for the formation of associative reward memories (Tye, Stuber, De Ridder, Bonci, & Janak, 2008) and BLA neurons respond to both reward and punishment in a pavlovian task with distinct underlying neuronal populations (Beyeler et al., 2016). Although no anatomical distinction seems to exist within the amygdala to separate ensembles that encode positive or negative valence stimuli, there are molecular markers for the encoding of valence. In particular, magnocellular Rspo2+ neurons are mainly activated by aversive stimuli while parvocellular Ppp1r1b+ neurons are mainly activated by appetitive stimuli (Kim, Pignatelli, Xu, Itohara, & Tonegawa, 2016), which suggests a molecular substrate for how the BLA could encode two antagonistic types of memory.

Classically, the vmPFC has been shown as a center for inhibitory control over drug-seeking (Moorman, James, McGlinchey, & Aston-Jones, 2015) and the dmPFC has been thought to drive drug-seeking behavior. This idea of dmPFC as a ‘go’ center and vmPFC as a ‘no-go’ center has an obvious parallel with fear conditioning studies, in which the dmPFC is related to fear expression and vmPFC is related to extinction learning (Giustino & Maren, 2015). However, a more nuanced view is necessary to understand mPFC function. For example, IL activation can either induce increase or decrease in drug-seeking (Koya et al., 2009; Peters, LaLumiere, & Kalivas, 2008), which could potentially be explained by a more complex time-dependent function of the vmPFC in the expression and extinction of cocaine-seeking (Van den Oever et al., 2013).

Cocaine self-administration results in enhancement of excitatory activity in the PL-to-NAc(core) pathway, while extinction results in an increase of excitation in the IL-to-NAc(shell) pathway. Both enhancements of excitation are mediated by synaptic maturation through the upregulation of AMPA receptors and optogenetic inhibition of the synaptic remodeling process results in a decrease of drug-seeking (Ma et al., 2014). While this study did not use ensemble-specific targeting, the evidence related to synaptic modeling with drug craving could fall in line with the proposed idea that the exertion of top-down processing of the mPFC depends on mature memory traces.

To direct assess ensemble function in addiction, a recent study has used the method of targeted recombination of active populations (TRAP) to specifically tag neurons that were naturally active during an alcohol self-administration task. The authors found that a small mPFC ensemble was necessary for cue-induced alcohol-seeking but not necessary for context-induced alcohol-seeking (i.e. when a salient cue was not present in the testing stage) (Visser et al., 2020). Furthermore, chemogenetic inactivation of the small alcohol-associated ensemble in the mPFC (6-7% of total neurons) led to a decrease in cue-induced alcohol-seeking, while inactivation of similarly sized sucrose-related ensemble did not lead to the same effect, which specifically demonstrates the effected of inactivation of the memory ensemble related to drug-seeking (Visser et al., 2020).

An important limitation of studies that use highly addictive drugs as rewards is the demonstration that cocaine-seeking becomes insensitive to devaluation after extensive training, which means that the behavior was not goal-directed, but rather habitual (Zapata, Minney, & Shippenberg, 2010). Therefore, in order to investigate goal-directed behavior, potentially a paradigm be used with two differences: (1) a less potent rewarding stimulus and (2) a task that has variance in the stimulus-response setups, such that an automatized pattern of behavior emerges from the mPFC.

5.2        DELAY-DISCOUNTING TASK (DDT)

Impulsivity can be defined as a premature action without foresight (Dalley, Everitt, & Robbins, 2011). The relationship between different brain regions and impulsivity has been studied for decades with paradigms such as the delay discounting task (DDT). In the DDT, the animal is trained to perform an action to receive a small reward (e.g. one food pellet) or wait for a few seconds and perform the same action to receive a large reward (e.g. five food pellets). One important methodological consideration is that the animal is unable to receive more rewards by performing a series of sequential short-term actions, therefore a time buffer needs to be put in place in between trials (Beckwith, 2017).

DDT is a powerful paradigm to assess a straightforward aspect of goal-direction. This paradigm has a reduction in confounding factors because both choices (short/small or long/large) have the same patterns of behavioral output, i.e. a waiting period followed by the same movement, which is different from the 5-choice serial-reaction time task, for example, where some responses require more movement than others. This is of extreme importance to study the mPFC because the dmPFC is also involved in motor planning (Euston & McNaughton, 2006). Therefore, in studying the waiting periods in the DDT, researchers can assess mPFC activity related to the choice of the animal, i.e. higher-order cognitive function without motor planning as a confounder.

In the DDT, pharmacological inactivation or neurotoxic lesions of the mPFC and disconnection between mPFC-BLA led to increased impulsivity (preference of short-term small reward over long-term big rewards) (Churchwell, Morris, Heurtelou, & Kesner, 2009; Gill, Castaneda, & Janak, 2010), This increase in impulsivity indicates a role in the mPFC of representing the outcome values of certain behavioral responses, which has later demonstrated with electrophysiological studies (Powell & Redish, 2016). In particular, dmPFC neurons are more active during the delay stage of the DDT and pharmacological inactivation of this brain region increases premature responses (Narayanan & Laubach, 2006).

Interestingly, electrophysiological recordings revealed an increase in PL activity during the delay period of a large reward, however, this increase was not observed when the animals were performing forced-choice trials (Sackett, Moschak, & Carelli, 2019). This parallels findings in habitual behavior research, where mPFC function is generally important for the learning stages of a new task, but once the contingencies of the task are learned, the mPFC becomes disengaged and behavior is driven mainly by subcortical structures, such as the dorsal striatum (Everitt & Robbins, 2016).

To date, no studies have investigated the mPFC using ensemble-specific targeting. A potential prolific field of research would investigate the development of ensemble maturation in the mPFC during the initial learning stages of an instrumental task up to when the actions of the animal become habitual, i.e. at which points of learning a complex task does the mPFC exert top-down control and at which points it does not.

6             CONCLUSIONS AND PERSPECTIVES FOR THE FIELD

In this review, we proposed a model of top-down and bottom-up as a hierarchy of sparse connectivity. This model, which places neuronal ensemble maturation as paramount for the function of top-down processing in the nervous system, could be a catalyzer for the integration of multiple lines of research, including negative and positive valence studies. In the mPFC, this entails that top-down control is necessary for goal-directed behavior, as evidenced from positive and negative valence paradigm studies, although the underlying process of maturation depends on the specific parameters of the task. Importantly, the presented evidence does not entail that the mPFC is the only region related to goal-directed behavior or that can exert top-down control. The role of the mPFC is an instantiation of the importance of association cortices in the integration of information from lower hierarchical structures. It is possible that ensemble maturation in other association cortices, such as the parietal cortex (Ivashkina, Gruzdeva, Roshchina, Toropova, & Anokhin, 2019), or even in a cortex-wide dispersed-ensemble fashion (Roy et al., 2019).

As for methodological considerations, the mPFC is a complex heteromodal associative cortex and its function not only depends on the task at hand, but it is also influenced by all previous experiential learnings of the animal throughout its life (Tse et al., 2007). Careful interpretation of results must be done to avoid problems such as the Euston-Cowen-McNaughton hassle, in which dmPFC activity encodes mainly motoric activity rather than cognitive function in a task where the animal is moving. Experimental design to study the mPFC would benefit from either: 1) using paradigms that have the same output response in both task contingencies (like the DDT) and; 2) use analysis of video recordings to try to unveil mPFC activity which is related to movement and mPFC activity related to the task (Mathis et al., 2018).

Furthermore, mPFC function is disrupted through stress, possibly mediated by the upregulation of dopaminergic and noradrenergic receptors, which would enhance calcium-cAMP pathways and ultimately would result in a reduction of mPFC activity and a subsequent reduction in connectivity of the ensembles (Datta & Arnsten, 2019). Therefore, standardization of protocols across labs and advances in technologies that minimize contact between researchers and animals to reduce stress, such as operant chambers connected to the home cage of the animal (Bruinsma et al., 2019), would be useful to reduce confounding effects and possibly contradictory findings in the field.

Another consideration is the fact that, under unchanging environmental conditions, learning becomes habitual, i.e. less dependent upon neocortical activity and more dependent on subcortical activity (Everitt & Robbins, 2016), indicating a loss of top-control of higher hierarchical structures. A useful experiment to elucidate the relationship between habits and ensemble development in the mPFC would be to use fos-Cre-GCaMP to visualize ensemble activity in the mPFC during the learning stages complex instrumental paradigm.

The main reason why ensemble-specific studies are powerful is their reduction of uncertainty in the interpretation of data. This is especially true for association cortices like the mPFC, which is integrated with motor, limbic and sensory cortices. In this brain region, the general manipulation of neurons via optogenetics or chemogenetics might induce downstream changes in neuronal activity which are unrelated to the task at hand. In specific targeting engrams, these confounding effects are reduced and the subsequent change in behavior can be interpreted more precisely.

Recent efforts have tried to unveil the intracellular and extracellular mechanisms of neuronal maturation. In terms of intracellular mechanisms, consolidation of fear memory requires BNDF upregulation in the PL, which leads to the increase in neuroligin 1 (NLGN1) and neuroligin 2 (NLGN2), important markers of synaptic maturation (Ye, Kapeller-Libermann, Travaglia, Inda, & Alberini, 2017). As for extracellular mechanisms, myelin plasticity seems to be important for fear learning: in transgenic animals that cannot produce myelin, there is a deficiency in remote fear memory recall. This phenotype can be partially rescued by the induction of myelin expression in the brain (Pan, Mayoral, Choi, Chan, & Kheirbek, 2020).

To conclude, a focus of future research on unveiling the specifics of ensemble maturation may yield fruitful results and allow the cross-communication between various lines of neuroscience research. Further research is still needed to assess: 1) how the valence and strength of a stimulus are related to ensemble maturation in the learning of different paradigms 2) what is the interplay between changes in vasculature, myelination, synaptic morphology and protein expression in the process of ensemble maturation.

7             REFERENCES

Alvarez, P., & Squire, L. R. (1994). Memory consolidation and the medial temporal lobe: A simple network model. Proceedings of the National Academy of Sciences of the United States of America, 91(15), 7041–7045. https://doi.org/10.1073/pnas.91.15.7041

An, B., Kim, J., Park, K., Lee, S., Song, S., & Choi, S. (2017). Amount of fear extinction changes its underlying mechanisms. ELife, 6. https://doi.org/10.7554/eLife.25224

Beckwith, S. W. (2017). The Role of the Medial Prefrontal Cortex in Delay Discounting.

Beyeler, A., Namburi, P., Glober, G. F., Simonnet, C., Calhoon, G. G., Conyers, G. F., … Tye, K. M. (2016). Divergent Routing of Positive and Negative Information from the Amygdala during Memory Retrieval. Neuron, 90(2), 348–361. https://doi.org/10.1016/j.neuron.2016.03.004

Bouton, M. E. (2004, September). Context and behavioral processes in extinction. Learning and Memory. Learn Mem. https://doi.org/10.1101/lm.78804

Bouton, M. E., & Bolles, R. C. (1979). Contextual control of the extinction of conditioned fear. Learning and Motivation, 10(4), 445–466. https://doi.org/10.1016/0023-9690(79)90057-2

Bruinsma, B., Terra, H., de Kloet, S. F., Luchicchi, A., Timmerman, A. J., Remmelink, E., … Mansvelder, H. D. (2019). An automated home-cage-based 5-choice serial reaction time task for rapid assessment of attention and impulsivity in rats. Psychopharmacology, 1–12. https://doi.org/10.1007/s00213-019-05189-0

Bukalo, O., Pinard, C. R., Silverstein, S., Brehm, C., Hartley, N. D., Whittle, N., … Holmes, A. (2015). Prefrontal inputs to the amygdala instruct fear extinction memory formation. Science Advances, 1(6), e1500251. https://doi.org/10.1126/sciadv.1500251

Carlén, M. (2017, October 27). What constitutes the prefrontal cortex? Science. American Association for the Advancement of Science. https://doi.org/10.1126/science.aan8868

Chen, B. T., Yau, H. J., Hatch, C., Kusumoto-Yoshida, I., Cho, S. L., Hopf, F. W., & Bonci, A. (2013). Rescuing cocaine-induced prefrontal cortex hypoactivity prevents compulsive cocaine seeking. Nature, 496(7445), 359–362. https://doi.org/10.1038/nature12024

Chiesa, A., Serretti, A., & Jakobsen, J. C. (2012). Mindfulness: Top-down or bottom-up emotion regulation strategy? https://doi.org/10.1016/j.cpr.2012.10.006

Cho, J. H., Deisseroth, K., & Bolshakov, V. Y. (2013). Synaptic encoding of fear extinction in mPFC-amygdala circuits. Neuron, 80(6), 1491–1507. https://doi.org/10.1016/j.neuron.2013.09.025

Churchwell, J. C., Morris, A. M., Heurtelou, N. M., & Kesner, R. P. (2009). Interactions Between the Prefrontal Cortex and Amygdala During Delay Discounting and Reversal. Behavioral Neuroscience, 123(6), 1185–1196. https://doi.org/10.1037/a0017734

Cohen, Y. E. (2009). Multimodal activity in the parietal cortex. Hear Res, 258(2), 100–105. https://doi.org/10.1016/j.heares.2009.01.011

Dalley, J. W., Everitt, B. J., & Robbins, T. W. (2011, February 24). Impulsivity, Compulsivity, and Top-Down Cognitive Control. Neuron. Cell Press. https://doi.org/10.1016/j.neuron.2011.01.020

Datta, D., & Arnsten, A. F. T. (2019). Loss of prefrontal cortical higher cognition with uncontrollable stress: Molecular mechanisms, changes with age, and relevance to treatment. Brain Sciences, 9(5), 1–16. https://doi.org/10.3390/brainsci9050113

de Kloet, E. R., de Kloet, S. F., de Kloet, C. S., & de Kloet, A. D. (2019). Top-down and bottom-up control of stress-coping. Journal of Neuroendocrinology, 31(3), 1–16. https://doi.org/10.1111/jne.12675

Duvarci, S., & Pare, D. (2014, June 4). Amygdala microcircuits controlling learned fear. Neuron. Cell Press. https://doi.org/10.1016/j.neuron.2014.04.042

Euston, D. R., & McNaughton, B. L. (2006). Apparent encoding of sequential context in rat medial prefrontal cortex is accounted for by behavioral variability. Journal of Neuroscience, 26(51), 13143–13155. https://doi.org/10.1523/JNEUROSCI.3803-06.2006

Everitt, B. J., & Robbins, T. W. (2016). Drug Addiction: Updating Actions to Habits to Compulsions Ten Years On. Annual Review of Psychology, 67(1), 23–50. https://doi.org/10.1146/annurev-psych-122414-033457

Eysenck, H. J. (1968). A theory of the incubation of anxiety/fear responses. Behaviour Research and Therapy, 6(3), 309–321. https://doi.org/10.1016/0005-7967(68)90064-8

Floresco, S. B., Block, A. E., & Tse, M. T. L. (2008). Inactivation of the medial prefrontal cortex of the rat impairs strategy set-shifting, but not reversal learning, using a novel, automated procedure. Behavioural Brain Research, 190(1), 85–96. https://doi.org/10.1016/j.bbr.2008.02.008

Fuster, J. M. (2001). Prefrontal Cortex. In International Encyclopedia of the Social & Behavioral Sciences (pp. 11969–11976). https://doi.org/10.1016/B0-08-043076-7/03465-3

Gale, G. D., Anagnostaras, S. G., Godsil, B. P., Mitchell, S., Nozawa, T., Sage, J. R., … Fanselow, M. S. (2004). Role of the Basolateral Amygdala in the Storage of Fear Memories across the Adult Lifetime of Rats. Journal of Neuroscience, 24(15), 3810–3815. https://doi.org/10.1523/JNEUROSCI.4100-03.2004

Gazzaley, A., & Nobre, A. C. (2012, February). Top-down modulation: Bridging selective attention and working memory. Trends in Cognitive Sciences. NIH Public Access. https://doi.org/10.1016/j.tics.2011.11.014

Gill, T. M., Castaneda, P. J., & Janak, P. H. (2010). Dissociable roles of the medial prefrontal cortex and nucleus accumbens core in goal-directed actions for differential reward magnitude. Cerebral Cortex, 20(12), 2884–2899. https://doi.org/10.1093/cercor/bhq036

Gilmartin, M. R., Balderston, N. L., & Helmstetter, F. J. (2014, August). Prefrontal cortical regulation of fear learning. Trends in Neurosciences. Elsevier Ltd. https://doi.org/10.1016/j.tins.2014.05.004

Giustino, T. F., & Maren, S. (2015). The role of the medial prefrontal cortex in the conditioning and extinction of fear. Frontiers in Behavioral Neuroscience, 9(NOVEMBER). https://doi.org/10.3389/fnbeh.2015.00298

Halladay, L. R., Kocharian, A., Piantadosi, P. T., Authement, M. E., Lieberman, A. G., Spitz, N. A., … Holmes, A. (2020). Prefrontal Regulation of Punished Ethanol Self-administration. Biological Psychiatry, (28), 1–12. https://doi.org/10.1016/j.biopsych.2019.10.030

Heidbreder, C. A., & Groenewegen, H. J. (2003). The medial prefrontal cortex in the rat: evidence for a dorso-ventral distinction based upon functional and anatomical characteristics. https://doi.org/10.1016/j.neubiorev.2003.09.003

Heroux, N. A., Robinson-Drummer, P. A., Sanders, H. R., Rosen, J. B., & Stanton, M. E. (2017). Differential involvement of the medial prefrontal cortex across variants of contextual fear conditioning. Learning and Memory, 24(8), 322–330. https://doi.org/10.1101/lm.045286.117

Herry, C., & Johansen, J. P. (2014, January 1). Encoding of fear learning and memory in distributed neuronal circuits. Nature Neuroscience. Nature Publishing Group. https://doi.org/10.1038/nn.3869

Ivashkina, O., Gruzdeva, A., Roshchina, M., Toropova, K., & Anokhin, K. (2019). Ex vivo and in vivo imaging of mouse parietal association cortex activity in episodes of cued fear memory formation and retrieval. BioRxiv. https://doi.org/10.1101/863589

Izquierdo, A., Wellman, C. L., & Holmes, A. (2006). Brief uncontrollable stress causes dendritic retraction in infralimbic cortex and resistance to fear extinction in mice. Journal of Neuroscience, 26(21), 5733–5738. https://doi.org/10.1523/JNEUROSCI.0474-06.2006

Kamigaki, T. (2019, March 1). Prefrontal circuit organization for executive control. Neuroscience Research. Elsevier Ireland Ltd. https://doi.org/10.1016/j.neures.2018.08.017

Kim, J., Pignatelli, M., Xu, S., Itohara, S., & Tonegawa, S. (2016). Antagonistic negative and positive neurons of the basolateral amygdala. Nature Neuroscience, 19(12), 1636–1646. https://doi.org/10.1038/nn.4414

Kitamura, T., & Inokuchi, K. (2014, February 19). Role of adult neurogenesis in hippocampal-cortical memory consolidation. Molecular Brain. BioMed Central Ltd. https://doi.org/10.1186/1756-6606-7-13

Kitamura, T., Ogawa, S. K., Roy, D. S., Okuyama, T., Morrissey, M. D., Smith, L. M., … Tonegawa, S. (2017). Engrams and circuits crucial for systems consolidation of a memory. Science, 356(6333), 73–78. https://doi.org/10.1126/science.aam6808

Klavir, O., Prigge, M., Sarel, A., Paz, R., & Yizhar, O. (2017). Manipulating fear associations via optogenetic modulation of amygdala inputs to prefrontal cortex. Nature Neuroscience, 20(6), 836–844. https://doi.org/10.1038/nn.4523

Knapska, E., & Maren, S. (2009). Reciprocal patterns of c-Fos expression in the medial prefrontal cortex and amygdala after extinction and renewal of conditioned fear. Learning and Memory, 16(8), 486–493. https://doi.org/10.1101/lm.1463909

Koya, E., Golden, S. A., Harvey, B. K., Guez-Barber, D. H., Berkow, A., Simmons, D. E., … Hope, B. T. (2009). Targeted disruption of cocaine-activated nucleus accumbens neurons prevents context-specific sensitization. Nature Neuroscience, 12(8), 1069–1073. https://doi.org/10.1038/nn.2364

Lee, Y. K., & Choi, J.-S. (2012). Inactivation of the Medial Prefrontal Cortex Interferes with the Expression But Not the Acquisition of Diff erential Fear Conditioning in Rats. Experimental Neurobiology, 21(1), 23–29. https://doi.org/10.5607/en.2012.21.1.23

Lonsdorf, T. B., Menz, M. M., Andreatta, M., Fullana, M. A., Golkar, A., Haaker, J., … Merz, C. J. (2017). Don’t fear ‘fear conditioning’: Methodological considerations for the design and analysis of studies on human fear acquisition, extinction, and return of fear. Neuroscience and Biobehavioral Reviews, 77, 247–285. https://doi.org/10.1016/j.neubiorev.2017.02.026

Ma, Y. Y., Lee, B. R., Wang, X., Guo, C., Liu, L., Cui, R., … Dong, Y. (2014). Bidirectional modulation of incubation of cocaine craving by silent synapse-based remodeling of prefrontal cortex to accumbens projections. Neuron, 83(6), 1453–1467. https://doi.org/10.1016/j.neuron.2014.08.023

Marek, R., Jin, J., Goode, T. D., Giustino, T. F., Wang, Q., Acca, G. M., … Sah, P. (2018). Hippocampus-driven feed-forward inhibition of the prefrontal cortex mediates relapse of extinguished fear. Nature Neuroscience, 21(3), 384–392. https://doi.org/10.1038/s41593-018-0073-9

Marek, R., Sun, Y., & Sah, P. (2019, January 1). Neural circuits for a top-down control of fear and extinction. Psychopharmacology. Springer Verlag. https://doi.org/10.1007/s00213-018-5033-2

Maren, S., Phan, K. L., & Liberzon, I. (2013, June). The contextual brain: Implications for fear conditioning, extinction and psychopathology. Nature Reviews Neuroscience. https://doi.org/10.1038/nrn3492

Maroun, M., Ioannides, P. J., Bergman, K. L., Kavushansky, A., Holmes, A., & Wellman, C. L. (2013). Fear extinction deficits following acute stress associate with increased spine density and dendritic retraction in basolateral amygdala neurons. European Journal of Neuroscience, 38(4), 2611–2620. https://doi.org/10.1111/ejn.12259

Mathis, A., Mamidanna, P., Cury, K. M., Abe, T., Murthy, V. N., Mathis, M. W., & Bethge, M. (2018). DeepLabCut: markerless pose estimation of user-defined body parts with deep learning. Nature Neuroscience, 21(9), 1281–1289. https://doi.org/10.1038/s41593-018-0209-y

Matos, M. R., Visser, E., Kramvis, I., van der Loo, R. J., Gebuis, T., Zalm, R., … van den Oever, M. C. (2019). Memory strength gates the involvement of a CREB-dependent cortical fear engram in remote memory. Nature Communications, 10(1), 1–11. https://doi.org/10.1038/s41467-019-10266-1

McClelland, J. L. (2013). Incorporating rapid neocortical learning of new schema-consistent information into complementary learning systems theory. Journal of Experimental Psychology: General, 142(4), 1190–1210. https://doi.org/10.1037/a0033812

Mcklveen, J. M., Myers, B., Herman, J. P., & Herman, J. P. (2015). The Medial Prefrontal Cortex: Coordinator of Autonomic, Neuroendocrine and Behavioural Responses to Stress. Journal of Neuroendocrinology, 27, 446–456. https://doi.org/10.1111/jne.12272

Miller, B. T., & D’Esposito, M. (2005). Searching for “the top” in top-down control. Neuron, 48(4), 535–538. https://doi.org/10.1016/j.neuron.2005.11.002

Moorman, D. E., James, M. H., McGlinchey, E. M., & Aston-Jones, G. (2015). Differential roles of medial prefrontal subregions in the regulation of drug seeking. Brain Research, 1628(Pt A), 130–146. https://doi.org/10.1016/j.brainres.2014.12.024

Morgan, M. A., Romanski, L. M., & LeDoux, J. E. (1993). Extinction of emotional learning: Contribution of medial prefrontal cortex. Neuroscience Letters, 163(1), 109–113. https://doi.org/10.1016/0304-3940(93)90241-C

Moscarello, J. M., & Maren, S. (2018). Flexibility in the face of fear: hippocampal–prefrontal regulation of fear and avoidance. Current Opinion in Behavioral Sciences, 19, 44–49. https://doi.org/10.1016/j.cobeha.2017.09.010

Narayanan, N. S., & Laubach, M. (2006). Top-Down Control of Motor Cortex Ensembles by Dorsomedial Prefrontal Cortex. Neuron, 52(5), 921–931. https://doi.org/10.1016/j.neuron.2006.10.021

Padilla-Coreano, N., Bolkan, S. S., Pierce, G. M., Blackman, D. R., Hardin, W. D., Garcia-Garcia, A. L., … Gordon, J. A. (2016). Direct Ventral Hippocampal-Prefrontal Input Is Required for Anxiety-Related Neural Activity and Behavior. Neuron, 89(4), 857–866. https://doi.org/10.1016/j.neuron.2016.01.011

Pan, S., Mayoral, S. R., Choi, H. S., Chan, J. R., & Kheirbek, M. A. (2020). Preservation of a remote fear memory requires new myelin formation. Nature Neuroscience, 23(4), 487–499. https://doi.org/10.1038/s41593-019-0582-1

Peters, J., LaLumiere, R. T., & Kalivas, P. W. (2008). Infralimbic prefrontal cortex is responsible for inhibiting cocaine seeking in extinguished rats. Journal of Neuroscience, 28(23), 6046–6053. https://doi.org/10.1523/JNEUROSCI.1045-08.2008

Phillips, R. G., & LeDoux, J. E. (1992). Differential Contribution of Amygdala and Hippocampus to Cued and Contextual Fear Conditioning. Behavioral Neuroscience, 106(2), 274–285. https://doi.org/10.1037/0735-7044.106.2.274

Pitkänen, A., Pikkarainen, M., Nurminen, N., & Ylinen, A. (2006). Reciprocal Connections between the Amygdala and the Hippocampal Formation, Perirhinal Cortex, and Postrhinal Cortex in Rat: A Review. Annals of the New York Academy of Sciences, 911(1), 369–391. https://doi.org/10.1111/j.1749-6632.2000.tb06738.x

Pitkänen, A., Savander, V., & LeDoux, J. E. (1997). Organization of intra-amygdaloid circuitries in the rat: An emerging framework for understanding functions of the amygdala. Trends in Neurosciences, 20(11), 517–523. https://doi.org/10.1016/S0166-2236(97)01125-9

Powell, N. J., & Redish, A. D. (2016). Representational changes of latent strategies in rat medial prefrontal cortex precede changes in behaviour. Nature Communications, 7. https://doi.org/10.1038/ncomms12830

Rauss, K., & Pourtois, G. (2013). What is bottom-up and what is top-down in predictive coding. Frontiers in Psychology, 4(MAY), 276. https://doi.org/10.3389/fpsyg.2013.00276

Robbins, T. W. (2007). Delay Discounting and Impulsive Choice. Neuroscience, 1–18.

Roland, P. E., Hanazawa, A., Undeman, C., Eriksson, D., Tompa, T., Nakamura, H., … Ahmed, B. (2006). Cortical feedback depolarization waves: A mechanism of top-down influence on early visual areas. Proceedings of the National Academy of Sciences of the United States of America, 103(33), 12586–12591. https://doi.org/10.1073/pnas.0604925103

Roy, D. S., Park, Y.-G., Ogawa, S. K., Cho, J. H., Choi, H., Kamensky, L., … Tonegawa, S. (2019). Brain-wide mapping of contextual fear memory engram ensembles supports the dispersed engram complex hypothesis. BioRxiv. https://doi.org/10.1101/668483

Rusu, S. I., & Pennartz, C. M. A. (2019, January 1). Learning, memory and consolidation mechanisms for behavioral control in hierarchically organized cortico-basal ganglia systems. Hippocampus. John Wiley and Sons Inc. https://doi.org/10.1002/hipo.23167

Sackett, D. A., Moschak, T. M., & Carelli, R. M. (2019). Prelimbic cortical neurons track preferred reward value and reflect impulsive choice during delay discounting behavior. Journal of Neuroscience, 39(16), 3108–3118. https://doi.org/10.1523/JNEUROSCI.2532-18.2019

Shomstein, S. (2012). Cognitive functions of the posterior parietal cortex: Top-down and bottom-up attentional control. Frontiers in Integrative Neuroscience, (JULY 2012). https://doi.org/10.3389/fnint.2012.00038

Stern, C. A. J., Gazarini, L., Vanvossen, A. C., Hames, M. S., & Bertoglio, L. J. (2014). Activity in prelimbic cortex subserves fear memory reconsolidation over time. Learning and Memory, 21(1), 14–20. https://doi.org/10.1101/lm.032631.113

Thompson, R. F. (2005). In Search of Memory Traces. Annual Review of Psychology, 56(1), 1–23. https://doi.org/10.1146/annurev.psych.56.091103.070239

Tozzi, A. (2015, October 22). Information processing in the CNS: a supramolecular chemistry? Cognitive Neurodynamics. Kluwer Academic Publishers. https://doi.org/10.1007/s11571-015-9337-1

Tse, D., Langston, R. F., Kakeyama, M., Bethus, I., Spooner, P. A., Wood, E. R., … Morris, R. G. M. (2007). Schemas and memory consolidation. Science, 316(5821), 76–82. https://doi.org/10.1126/science.1135935

Tye, K. M., Stuber, G. D., De Ridder, B., Bonci, A., & Janak, P. H. (2008). Rapid strengthening of thalamo-amygdala synapses mediates cue-reward learning. Nature, 453(7199), 1253–1257. https://doi.org/10.1038/nature06963

Uylings, H. B. M., & Van Eden, C. G. (1991). Chapter 3 Qualitative and quantitative comparison of the prefrontal cortex in rat and in primates, including humans. Progress in Brain Research, 85(C), 31–62. https://doi.org/10.1016/S0079-6123(08)62675-8

Van den Oever, M. C., Rotaru, D. C., Heinsbroek, J. A., Gouwenberg, Y., Deisseroth, K., Stuber, G. D., … Smit, A. B. (2013). Ventromedial prefrontal cortex pyramidal cells have a temporal dynamic role in recall and extinction of cocaine-associated memory. Journal of Neuroscience, 33(46), 18225–18233. https://doi.org/10.1523/JNEUROSCI.2412-13.2013

Verschure, P. F. M. J., Pennartz, C. M. A., & Pezzulo, G. (2014). The why, what, where, when and how of goal-directed choice: Neuronal and computational principles. Philosophical Transactions of the Royal Society B: Biological Sciences, 369(1655). https://doi.org/10.1098/rstb.2013.0483

Visser, E., Matos, M. R., Van Der Loo, R. J., Marchant, N. J., De Vries, T. J., Smit, A. B., & Van Den Oever, M. C. (2020). A persistent alcohol cue memory trace drives relapse to alcohol seeking after prolonged abstinence. Retrieved from http://advances.sciencemag.org/

White, M. G., Panicker, M., Mu, C., Carter, A. M., Roberts, B. M., Dharmasri, P. A., & Mathur, B. N. (2018). Anterior Cingulate Cortex Input to the Claustrum Is Required for Top-Down Action Control. Cell Reports, 22(1), 84–95. https://doi.org/10.1016/j.celrep.2017.12.023

Ye, X., Kapeller-Libermann, D., Travaglia, A., Inda, M. C., & Alberini, C. M. (2017). Direct dorsal hippocampal-prelimbic cortex connections strengthen fear memories. Nature Neuroscience, 20(1), 52–61. https://doi.org/10.1038/nn.4443

Zapata, A., Minney, V. L., & Shippenberg, T. S. (2010). Shift from goal-directed to habitual cocaine seeking after prolonged experience in rats. Journal of Neuroscience, 30(46), 15457–15463. https://doi.org/10.1523/JNEUROSCI.4072-10.2010

Zelikowsky, M., Hersman, S., Chawla, M. K., Barnes, C. A., & Fanselow, M. S. (2014). Neuronal ensembles in amygdala, hippocampus, and prefrontal cortex track differential components of contextual fear. Journal of Neuroscience, 34(25), 8462–8466. https://doi.org/10.1523/JNEUROSCI.3624-13.2014

Zelinski, E. L., Hong, N. S., Tyndall, A. V., Halsall, B., & McDonald, R. J. (2010). Prefrontal cortical contributions during discriminative fear conditioning, extinction, and spontaneous recovery in rats. Experimental Brain Research, 203(2), 285–297. https://doi.org/10.1007/s00221-010-2228-0

Zwosta, K., Ruge, H., Wolfensteller, U., Heekeren, H. R., Noonan, M. P., & Farrer, C. (2015). Neural mechanisms of goal-directed behavior: outcome-based response selection is associated with increased functional coupling of the angular gyrus. https://doi.org/10.3389/fnhum.2015.00180

Note: This essay was written during my Master’s degree for the course Neurophilosophy and Ethics. I got a 7.5 (B+) and thought I would share it.

Human senses are incredibly limited and fallible. We only experience a small fraction of all the existing sounds and light frequencies (for instance, we see between 400 nm to 700 nm out of an infinite electromagnetic spectrum)¹, but even more fundamentally than that, the perception of a ‘true reality’ is inherently impossible because all information that arrives into our consciousness is inexorably multivariate.²

Figure 1. The yellow stickers on the left panel are the exact same shade of grey as the blue stickers on the right panel. Due to the inverse problem, our brain makes a probabilistic decision about what color to see based on previous experiences. 

The fundamental ambiguity of our perception of reality, the so-called ‘inverse problem’ was first described in detail by Hermann von Helmhortz³, and it can be exemplified by our sense of vision: what we perceive as ‘color’ comprises three elements: 1) the light characteristics that arrive from a light source; 2) the way that they bounce off of an object and; 3) and the air between the object and our eyes. This means that what we perceive as color has three indistinguishable variables: illumination (which is intrinsic to the light source), reflectance (how the object interacts with light) and transmittance (the influence of the atmosphere between object and eye). As a consequence, our brains need to make judgments about ‘which color to see’, which can result in interesting optical illusions (Figure 1)². Similar problems of multivariability exist for all of our senses, which means that reality is fundamentally inacessible to our brains.

This seems counterintuitive since humans are arguably the most successful species that have ever existed: we are on the top of our food chain, we lack natural predators and we are capable of inhabiting virtually every ecosystem on earth⁴ — which must entail that we have been able to adequately navigate our environment. How is it possible that we can appropriately change our behaviors based on our perception of reality if reality itself cannot be assessed by our senses? One possible answer is that we have a Bayesian brain — we make decisions on experiential information and construct our beliefs and decide courses of action based on the probability that they are correct.

Figure 2. A weighing scale can be used as a metaphor for the Bayesian brain.

A metaphor that could explain the principle behind the Bayesian brain is the following: Imagine an old school weighing scale, where objects can be put on each side to verify which one is the heaviest (Figure 2). This can serve as a visual metaphor for the way our brains make decisions based on probabilities: we attribute ‘weight’ to pieces of evidence that we encounter over our lifetime and whichever side of a belief has the most weight, ‘wins’ – becoming the thing that we believe.

One could explain this principle with an extremely simplified binary example: the belief “I like or I do not like bananas” are on the two ends of a scale. Our brain makes the decision about which one to believe (and then act upon) by weighing the evidence for “I like bananas” (maybe I have eaten hundreds of bananas and I always enjoyed them) and the opposing evidence for “I don’t like bananas” (maybe I once ate a banana with a worm in it) against each other. Each ‘pebble of evidence’ has its own weight and whichever side of the scale has the most weight, wins over the other side and becomes the belief we hold. Every new experience with bananas gives another ‘pebble of evidence’ in the scale — which means that we are constantly updating our beliefs based on new evidence, and then changing our behavior based on that updated scale.

Some researchers believe that the principle behind this analogy applies to how the entire brain works!^5–8 Every sense that we have (smell, touch, vision) do not represent reality: they represent a set of beliefs that we have obtained through experience in the natural world, which we use as a way to make decisions. A simple example of the role of experience in our perception is the fact that humans tend to interpret nearly-vertical lines as completely vertical and nearly-horizontal lines as completely horizontal – because these are the patterns most commonly observed in nature.⁹

Figure 3. The Necker cube is an example of bistable perception.

This metaphor has many great features: as with a normal scale, it can be in equilibrium — e.g. for the belief ‘I like or I do not like bananas’, if your experiential evidence is equivalent for both sides, you may come to the conclusion ‘I don’t mind bananas, I do not like them but I do not dislike them either’. As applied to our sense of vision, one could imagine this state of ‘belief equilibrium’ exemplified by the Necker cube, in which we have a bistable perception about which face is closest to us (either the lower-left or the upper-right)¹⁰ – which could be analogous to the scale slightly wobbling from left to right depending on which part of the figure we choose to focus on.

Another good feature of the metaphor is that the distinction between the number of pieces of evidence and their ‘value’ makes intuitive sense: if you have 100 pebbles that weigh 1 g each and one pebble that weighs 1000 g, should you place them on opposite sides of the scale, the single heavier pebble will win. Similarly, 100 good experiences may be outweighed by one horrible experience in determining one’s belief about the fruit (for instance, eating a banana with a worm in it, in the decision ‘I like or dislike bananas’).

Because the metaphor is quite simple, it can be also extended to explain many facets of the human experience and how they relate to this probabilistic nature of the Bayesian brain:

 1) Humans have intrinsic fears of certain animals, such as snakes and spiders, developed throughout the mammalian evolutionary history.¹¹ For the belief “I like or I do not like spiders”, we could visualize that the scale is naturally tipped to the dislike side — meaning that, even without any experiential evidence, humans have an intrinsic fear of arachnids.

2) Humans have a salience bias, meaning that emotionally evocative information is more persuasive than neutral ones. For instance, when deciding for the belief  ‘Planes are safe or unsafe’, one may conclude the latter based on shocking imagery they once saw on TV of a plane crash — even if statistics show that planes are objectively one of the safest modes of transportation.¹² In the context of the metaphor, one may visualize that the evidence pebbles have an emotional component that gives them extra weight.

3) Humans tend to make judgments based on more recent information – which is a phenomenon known as the ‘availability heuristic’.¹³ For example, when deciding ‘Elevators are safe or unsafe’, one’s judgment might be severely skewed to the ‘unsafe’ side if they saw a coworker get stuck in an elevator two days before. In the scale metaphor, one could imagine that the pebbles of evidence wear out and become lighter over time, exerting less influence over the belief scale.

As far as disadvantages, the metaphor is fairly abstract in content, which means that it can be used to explain essentially anything – as exemplified in this article, in which I use it to explain food preferences, visual perception, and cognitive biases. This could be considered a strong weakness because it is often unclear how the individual elements of the metaphor relate to reality. The lack of clarity may induce misunderstandings among laypeople: do humans have millions of scales inside their heads, each one making one individual decision?

The metaphor also attempts to reduce all complexities of learning and decision-making into single inanimate objects. This not only can be viewed as an oversimplification of what occurs in the brain, but it implicates a hard deterministic interpretation of free will¹⁴: by stating that beliefs are explained by previous experiences and biological predispositions under the gaze of a purely physical process (placing weights on top of scales), the imagery strongly hints to a fully deterministic universe which does not leave space for the sense of one’s agency in modifying their beliefs.

Another strong weakness of the metaphor is that it does not account for the fact that beliefs are interdependent¹⁵: ‘I like or dislike bananas’ is not only dependent on one’s past experience, but also dependents on their current beliefs and feelings, such as ‘I like or dislike fruits in general’, ‘I am really hungry right now’ or ‘I am really tired and I can’t even think about food’. Maybe the metaphor could be extended to include that each ‘belief scale’ is converted into a pebble of evidence into another belief scale, or perhaps you place scales on top of scales to represent this interdependency – but it then becomes convoluted and unclear as an explanatory device.

It is worth noting that the metaphor can be vague in meaning because the hypothesis itself is also fairly abstract: the Bayesian Brain hypothesis has yielded many criticisms, including one of being unscientific¹⁶. The statement “The brain is performing probabilistic decisions like a computer” does not provide a mechanism by which the computations are performed and it is therefore unfalsifiable. Under a strict Popperian view of science, this would mean that the hypothesis is unscientific and therefore not worthy of any merit.

Science still does not have a definitive answer about fundamental processes underlying our decision making, and the Bayesian brain is just one hypothesis that tries to explain the disconnect between ‘why we cannot perceive reality’ and ‘how we operate in the environment’. The Bayesian brain has been put in vogue in the last decade, especially by computer scientists who use machine learning techniques that rely heavily on principles of Bayesian statistics¹⁷. However, just because computers use probabilistic statistics to make decisions, it does not follow that humans necessarily use the same mechanism – this would be an inductive fallacy of faulty generalization. The proponents of the Bayesian Brain hypothesis still need to demonstrate the existence of this underlying mechanism in humans.

Science communicators must be cautious in the way they explain science in popular media to properly separate what is established science from speculative hypotheses and personal opinions, especially in the field of Neuroscience, in which a lot of mysticism persists in the zeitgeist¹⁸. However, it also means that some of the most burning questions one could ask about the world are not scientific: no experiment could ever be devised to experimentally test what the true nature of reality is¹⁹. This begs the question: should we only focus on science to understand reality and forget about unfalsifiable hypotheses or should we be open to other forms of explaining how the world works?

To conclude, the use of metaphors is a commonly used device for explanations, because our brains have higher developed visual and language cortices, it is often easier to explain concepts in terms of vivid imagery and stories^20,21; however, we need to be careful about the precision of our metaphors and possible misunderstandings they might generate. The scale metaphor provides an easy and generalizable way to understand the process of decision-making in the human brain, but it fails to account for the interdependence of beliefs and it diminishes one’s agency by appealing to purely physical processes.

References

1.        Dale Purves, George Augustine, David Fitzpatrick, William Hall, Anthony-Samuel Lamantia, L. W. Neuroscience. 5th Edition. (Sinauer Associates, 2012).

2.        Purves, D., & Lotto, R. B. Why we see what we do redux: a wholly empirical theory of vision. Choice Reviews Online 48, (Sinauer Associates., 2011).

3.        Lübbig, H. & Helmholtz, H. von. The inverse problem : symposium ad memoriam Hermann von Helmholtz. Research, measurement, approval, xii, 225 p. (1995).

4.        Darimont, C. T., Fox, C. H., Bryan, H. M. & Reimchen, T. E. The unique ecology of human predators. Science (80-. ). 349, 858–860 (2015).

5.        Yuille, A. & Kersten, D. Vision as Bayesian inference: analysis by synthesis? Trends Cogn. Sci. 10, 301–308 (2006).

6.        Wolpert, D. M. Probabilistic models in human sensorimotor control. Hum. Mov. Sci. 26, 511–524 (2007).

7.        Pantelis, P. C. et al. Inferring the intentional states of autonomous virtual agents. Cognition 130, 360–379 (2014).

8.        Houlsby, N. M. T. et al. Cognitive tomography reveals complex, task-independent mental representations. Curr. Biol. 23, 2169–2175 (2013).

9.        Girshick, A. R., Landy, M. S. & Simoncelli, E. P. Cardinal rules: Visual orientation perception reflects knowledge of environmental statistics. Nat. Neurosci. 14, 926–932 (2011).

10.      Wernery, J. et al. Temporal processing in bistable perception of the Necker cube. Perception 44, 157–168 (2015).

11.      Öhman, A. & Mineka, S. Fears, phobias, and preparedness: Toward an evolved module of fear and fear learning. Psychol. Rev. 108, 483–522 (2001).

12.      Slovic, P. The perception of risk. Sci. Mak. a Differ. One Hundred Eminent Behav. Brain Sci. Talk about their Most Important Contrib. 236, 179–182 (2016).

13.      Schwarz, N. et al. Ease of Retrieval as Information: Another Look at the Availability Heuristic. J. Pers. Soc. Psychol. 61, 195–202 (1991).

14.      Scheer, R. K. Free Will. Philosophical Investigations 13, (Free Press, 1990).

15.      Glore, G. L. & Gasper, K. Feeling is believing: Some affective influences on belief. Emot. Beliefs 10–44 (2010). doi:10.1017/cbo9780511659904.002

16.      Horgan, J. Are Brains Bayesian? Scientific American (2017). Available at: https://blogs.scientificamerican.com/cross-check/are-brains-bayesian/. (Accessed: 27th January 2020)

17.      Bobrov, P. D., Korshakov, A. V., Roshchin, V. I. & Frolov, A. A. [Bayesian classifier for brain-computer interface based on mental representation of movements]. Zh. Vyssh. Nerv. Deiat. Im. I P Pavlova 62, 89–99 (2012).

18.      O’Connor, C., Rees, G. & Joffe, H. Neuroscience in the Public Sphere. Neuron 74, 220–226 (2012).

19.      Jackendoff, R. The Problem of Reality. Noûs 25, 411 (1991).

20.      Exhibit, S., Batsak, B. & Orlov, O. Visual metaphor comprehension : an fMRI study. (2018).

21.      Zadbood, A., Chen, J., Leong, Y. C., Norman, K. A. & Hasson, U. How We Transmit Memories to Other Brains: Constructing Shared Neural Representations Via Communication. Cereb. Cortex 27, 4988–5000 (2017).

Note: This essay was written during my Master’s degree for the course Methods in Behavioral Neuroscience. I got a 7.5 (B+) for it, so I thought I would share.

Abstract

The processes that underlie aversive memory formation have long been elusive, in part because of the complexity of the brain-wide, sparse network of cortical regions that influence fear behavior, which includes the amygdala, hippocampus and mPFC. The mPFC is thought to provide top-down control over behavioral output, but the specifics of the microcircuitry and role of specific neuronal subtypes that are involved in this process remain poorly understood. The goal of this mini-review paper is to obtain a better insight into the role of the mPFC in fear conditioning by providing an overview of papers that have used state-of-the-art technology and provide future perspectives in the field.

What is the PFC?

The prefrontal cortex is the association cortex of the frontal lobe (Fuster, 2001). The PFC is involved in working memory, attention and decision making, cognitive processes collectively referred to as “executive functions” (Domenech & Koechlin, 2015). These higher-level cognitive functions of the PFC derive from its unique anatomy – it has extensive afferent and efferent connections to many brain regions, including the hippocampus, amygdala, basal ganglia, motor and sensory cortices (Gabbott et al., 2005). The PFC is therefore thought to be the ‘goal-directed’ cortex because it exerts top-down control over limbic, sensory and motor information and allows the organism to perform the most optimal behavior to the specific environmental conditions while inhibiting alternative actions primed by these bottom-up systems (Kamigaki, 2019).

Much of our current understanding comes from animal models, particularly mice. It may initially seem counterintuitive that the research on the rodent PFC could be useful to humans, since this brain region comprises 30% of the human cortical area, it is the last brain region to mature in adults and has specific anatomical regions not analogous to any other mammal (Neubert et al., 2014). While it is true that the murine brain is much simpler and has significant anatomical differences to the human PFC, e.g. a lack the granular layer IV in the cortex, it is now widely accepted that the mouse PFC has functional similarities: they can learn tasks that require attention (e.g. 5CSRTT), working memory (e.g. delayed matching-to-position task) and decision making (e.g. delayed reward paradigm).

The rodent PFC is broadly divided into three regions: ventral, lateral and medial. The medial prefrontal cortex (mPFC) has been extensively studied and found to be associated with many processes, including executive functions, recent and long memory formation (Euston, Gruber, & McNaughton, 2012). The mPFC can be broadly subdivided into a dorsomedial portion (dmPFC) and a ventromedial portion (vmPFC) which are differently connected to other brain regions: the vmPFC is more highly connected to limbic structures while the dmPFC is more highly connected to sensory and motor regions. The dmPFC comprises the anterior cingulate (AC) and the dorsal-most part of the prelimbic cortex (PL). The vmPFC comprises the infralimbic cortex (IL) and the ventral-most part of the prelimbic cortex. Because there is no clear division between dmPFC and vmPFC, only a gradient of projections to cortical and limbic structures in the PL (Gabbott et al., 2005), this review will mention the specific mPFC subregion whenever possible.

The goal-orienting nature of the mPFC is evolutionarily important because of the unpredictability of nature – while instincts can be developed for intrinsically stereotypical cues, much of the environment is highly context-dependent, and the organism must learn over its lifetime good or bad associations between environmental stimuli, behavior and outcome results. Fear conditioning has been extensively used to study long-term aversive memory and how sensory input ultimately generate adaptive behavior (Herry & Johansen, 2014). This paradigm has many advantages over other aversive memory paradigms, principally because it allows for a minimal amount of training – one session with a duration of a few minutes will promote permanent memory changes.

The goal of this mini-review paper is to obtain a better insight into the role of the mPFC in the orchestration of behavior in response to aversive stimuli, addressing current controversies in the published literature and indicating potential future research directions that could address the current gaps in knowledge.

The role of the mPFC in fear memory

Fear conditioning involves the association of a neutral stimulus contingently paired with an aversive stimulus. In this paradigm, a conditioned stimulus (e.g. an auditory tone) is paired with an unconditioned stimulus (e.g. a foot shock). The conditioned auditory stimulus and unconditioned shock stimulus enter the amygdala at the level of the basal and lateral nuclei of the amygdala (collectively known as BLA), where they are biologically paired. The BLA is connected to the central nucleus (CeA), which is the major output system of the amygdala, which then promotes the conditioned response. The neuronal connections that underlie a memory are named memory engrams (Josselyn, 2010). The level of fear is measured from the stereotypical behavioral output – i.e. how much time the animal spends on the freezing state (Jacobs, Cushman, & Fanselow, 2010), and manipulations in the memory engram can be performed with techniques such as optogenetics, chemogenetics or activity-dependent labeling.

The mPFC is highly interconnected with the amygdala, mainly with the BLA. The IL is thought to mediate extinction by increasing the activity of GABAergic intercalated cells, which decreases the activity of the CeA. The PL is thought to increase the activity of glutamatergic neurons in the BLA, increasing the output of the CeA and promoting the fear response (Marek, Sun, & Sah, 2019). However, the PL and IL are not as independent as this simplified model might suggest. There are glutamatergic neurons that project from PL to IL, and specific excitation of these enhances fear extinction (Marek et al., 2018). Davis et al. (2017) showed that chemogenetic inhibition of BLA parvalbumin-positive (PV) interneurons lead to an increase in freezing behavior after extinction learning. Using active dependent labeling, the authors found that this disinhibition in the BLA leads to the reactivation of engram neurons in the mPFC – indicating a functional role of BLA PV interneurons in extinction. In the same study, the authors found that the IL had more projections to BLA PV interneurons compared to PL projections and the BLA had more projections to the PL – establishing a reciprocal loop between mPFC and BLA.

The process of extinction, i.e. weakening the association between CS-US upon repeated exposures of the same context without the shock, is mediated by the IL. Extinction does not erase the original fear memory, but rather creates a new memory (An et al., 2017). Extinction learning can be enhanced by optogenetically stimulating the IL  and impaired by chemogenetically silencing this brain region (Bloodgood et al., 2018). This leads to an interesting question of how output behavior could be mediated with two contradictory fear memories. Davis et al. (2017) used in vivo electrophysiology to record monosynaptic BLA PV interneurons projecting to the mPFC and found that fear conditioning leads to an increase in the 3-6 Hz range, silencing BLA PV interneurons decreased the 6-12 Hz range and the ratio of these two frequencies in the mPFC was correlated with the amount of freezing after extinction, but not before. This shift in network activity may be the neuronal correlate of the competition between fear memory and extinction memory.

The mPFC is also highly interconnected to the hippocampus, which is thought to provide contextual information about the fear memory. Rajasethupathy et al. (2015)used optogenetics to stimulate monosynaptic projections from the hippocampus to the AC after fear conditioning and found that mice would increase freezing behavior in an unrelated context. This effect was thought to be specific to the fear memory because, after extinction, the activation of this pathway no longer increased fear behavior. After placing the mice in a new context, the fear memory was reinstated, and the optogenetic activation again increased the fear memory. Kitamura et al. (2017) used a combination of activity-dependent cell-labeling, optogenetics and miniscopes to assess the activity of projections from the hippocampus to the PFC related to the fear memory. Inhibition of the engram had no significant effects on early fear memory but disrupted remote fear memory. The engram neurons in the mPFC were active in the original context in which the animals had received the shock 13 days after the original fear conditioning, but were silent 2 days later – i.e. the memory was not retrievable by environmental cues but could be activated with optogenetics. On the other hand, the hippocampal engram was active 2 days after fear conditioning, but silent 13 days later. The same authors hypothesize that the fear memory is initially encoded simultaneously in the mPFC, amygdala and hippocampus, but the engram undergoes maturation at different time points in the mPFC and hippocampus (Figure 1).

Contradictions in current findings

The mPFC has a great degree of complexity and interconnection to other brain regions: it comprises around 90% glutamatergic neurons, 10% GABAergic neurons and receives projections from several neuromodulator centers (serotonin from Ralphe nucleus, dopamine from the VTA, acetylcholine from the basal forebrain and noradrenaline from the locus coeruleus) (Dembrow & Johnston, 2014). Therefore, a unitary function for the mPFC cannot be expected – the role of this brain region will vary depending on specifics of the task at hand, sensory inputs, previous memory associations and internal needs of the animal. For example, the inactivation of PL has been found to have a number of contradictory effects in aversively motivated procedures – increasing fear memory, decreasing fear memory or having no effect at all (Sharpe & Killcross, 2018). The authors attribute those differences to differences in methodology – for instance, the duration of training and intertraining intervals, which may alter the learning association of the context with the aversive stimuli. Inactivation of the mPFC has been found to have no effect on acquisition, consolidation or retention of fear memory using Pavlovian condition, but it was required for acquisition, consolidation and retention when these three aspects of learning were temporally spaced (Heroux et al., 2017).

These contradictory effects could also be explained by the variability in engram targeting. When inactivating a brain region with chemogenetics/optogenetics, it is possible that an engram could or could not be targeted. The advance of more precise techniques for neuronal manipulation allows for the unraveling of the exact role that each cell type plays in the fear engram. For instance, Adhikari et al. (2015) found that optogenetically inhibiting projections from the vmPFC to the amygdala decreased the fear memory, while inhibiting dmPFC-ITC projections showed a decrease in fear behavior only in cued fear-extinction. The authors found that inactivating this specific vmPFC population leads to the suppression of fear memory, but not bulk activation of the entire vmPFC – which suggests a high variance of cell types in this region that may counteract the effects of the vmPFC-amygdala pathway. It is therefore of extreme importance to understand the mechanisms that make the memory engram active or silenced. Some of the possible mechanisms are: epigenetic changes in histone subunit (Zovkic et al., 2014), upregulation of AMPA receptors in engram neurons (Arruda-Carvalho & Clem, 2014), ratio of oscillatory activity between mPFC and BLA in the 3-6 Hz range over the 6-12 Hz range (Davis et al., 2017).

To conclude, the mPFC plays an important role in fear memory – remaining as a silent engram during early consolidation but becoming involved after late consolidation. Engram and cell-specific studies have been shown to be of incredible value, allowing for the modulation of neurons related to the substrate of fear memories and answering important questions raised by contradictory research done in the past. Current important questions remain: how are the neuronal substrates different for different types of memory (e.g. episodic/semantic)? How much of this knowledge applies to paradigms of different valences (either less stressful aversive memories or even positive memory paradigms)? And finally, how much of this knowledge is transferable to humans, based not only on the anatomical differences between humans and rodents but also the ecological validity of these paradigms? Future research could benefit from the use of more complex paradigms that use both aversive and appetitive stimuli depending on different cues and using miniscopes to assess how the engrams are differently formed and how much they overlap. The advance of the experimental toolbox for other mammals would also be of extreme importance since most of the current research on this topic uses mice and some of the findings could be species-specific. Replicating these findings using cell-specific and engram targeting in non-human primates would be of extreme relevance for the translational aspect of memory research.

References

Adhikari, A., Lerner, T. N., Finkelstein, J., Pak, S., Jennings, J. H., Davidson, T. J., … Deisseroth, K. (2015). Basomedial amygdala mediates top-down control of anxiety and fear. Nature, 527(7577), 179–185. https://doi.org/10.1038/nature15698

An, B., Kim, J., Park, K., Lee, S., Song, S., & Choi, S. (2017). Amount of fear extinction changes its underlying mechanisms. ELife, 6. https://doi.org/10.7554/eLife.25224

Arruda-Carvalho, M., & Clem, R. L. (2014). Pathway-selective adjustment of prefrontal-amygdala transmission during fear encoding. Journal of Neuroscience, 34(47), 15601–15609. https://doi.org/10.1523/JNEUROSCI.2664-14.2014

Bloodgood, D. W., Sugam, J. A., Holmes, A., & Kash, T. L. (2018). Fear extinction requires infralimbic cortex projections to the basolateral amygdala. Translational Psychiatry, 8(1), 60. https://doi.org/10.1038/s41398-018-0106-x

Davis, P., Zaki, Y., Maguire, J., & Reijmers, L. G. (2017). Cellular and oscillatory substrates of fear extinction learning. Nature Neuroscience, 20(11), 1624–1633. https://doi.org/10.1038/nn.4651

Dembrow, N., & Johnston, D. (2014). Subcircuit-specific neuromodulation in the prefrontal cortex. Frontiers in Neural Circuits, 8(JUNE), 1–9. https://doi.org/10.3389/fncir.2014.00054

Domenech, P., & Koechlin, E. (2015). Executive control and decision-making in the prefrontal cortex. Current Opinion in Behavioral Sciences, 1, 101–106. https://doi.org/10.1016/j.cobeha.2014.10.007

Euston, D. R., Gruber, A. J., & McNaughton, B. L. (2012). The Role of Medial Prefrontal Cortex in Memory and Decision Making. Neuron, 76(6), 1057–1070. https://doi.org/10.1016/j.neuron.2012.12.002

Fuster, J. M. (2001). The prefrontal cortex – An update: Time is of the essence. Neuron (Vol. 30). https://doi.org/10.1016/S0896-6273(01)00285-9

Gabbott, P. L. A., Warner, T. A., Jays, P. R. L., Salway, P., & Busby, S. J. (2005). Prefrontal cortex in the rat: Projections to subcortical autonomic, motor, and limbic centers. Journal of Comparative Neurology, 492(2), 145–177. https://doi.org/10.1002/cne.20738

Heroux, N. A., Robinson-Drummer, P. A., Sanders, H. R., Rosen, J. B., & Stanton, M. E. (2017). Differential involvement of the medial prefrontal cortex across variants of contextual fear conditioning. Learning and Memory, 24(8), 322–330. https://doi.org/10.1101/lm.045286.117

Herry, C., & Johansen, J. P. (2014). Encoding of fear learning and memory in distributed neuronal circuits. Nature Neuroscience, 17(12), 1644–1654. https://doi.org/10.1038/nn.3869

Jacobs, N. S., Cushman, J. D., & Fanselow, M. S. (2010). The accurate measurement of fear memory in Pavlovian conditioning: Resolving the baseline issue. Journal of Neuroscience Methods, 190(2), 235–239. https://doi.org/10.1016/j.jneumeth.2010.04.029

Josselyn, S. A. (2010). Continuing the search for the engram: Examining the mechanism of fear memories. Journal of Psychiatry and Neuroscience, 35(4), 221–228. https://doi.org/10.1503/jpn.100015

Kamigaki, T. (2019, March 1). Prefrontal circuit organization for executive control. Neuroscience Research. Elsevier Ireland Ltd. https://doi.org/10.1016/j.neures.2018.08.017

Kitamura, T., Ogawa, S. K., Roy, D. S., Okuyama, T., Morrissey, M. D., Smith, L. M., … Tonegawa, S. (2017). Engrams and circuits crucial for systems consolidation of a memory. Science, 356(6333), 73–78. https://doi.org/10.1126/science.aam6808

Marek, R., Sun, Y., & Sah, P. (2019). Neural circuits for a top-down control of fear and extinction. Psychopharmacology, 236(1), 313–320. https://doi.org/10.1007/s00213-018-5033-2

Marek, R., Xu, L., Sullivan, R. K. P., & Sah, P. (2018). Excitatory connections between the prelimbic and infralimbic medial prefrontal cortex show a role for the prelimbic cortex in fear extinction. Nature Neuroscience, 21(5), 654–658. https://doi.org/10.1038/s41593-018-0137-x

Neubert, F. X., Mars, R. B., Thomas, A. G., Sallet, J., & Rushworth, M. F. S. (2014). Comparison of Human Ventral Frontal Cortex Areas for Cognitive Control and Language with Areas in Monkey Frontal Cortex. Neuron, 81(3), 700–713. https://doi.org/10.1016/j.neuron.2013.11.012

Rajasethupathy, P., Sankaran, S., Marshel, J. H., Kim, C. K., Ferenczi, E., Lee, S. Y., … Deisseroth, K. (2015). Projections from neocortex mediate top-down control of memory retrieval. Nature, 526(7575), 653–659. https://doi.org/10.1038/nature15389

Sharpe, M. J., & Killcross, S. (2018). Modulation of attention and action in the medial prefrontal cortex of rats. Psychological Review, 125(5), 822–843. https://doi.org/10.1037/rev0000118

Tonegawa, S., Morrissey, M. D., & Kitamura, T. (2018). The role of engram cells in the systems consolidation of memory. Nature Reviews Neuroscience, 19(8), 485–498. https://doi.org/10.1038/s41583-018-0031-2

Zovkic, I. B., Paulukaitis, B. S., Day, J. J., Etikala, D. M., & Sweatt, J. D. (2014). Histone H2A.Z subunit exchange controls consolidation of recent and remote memory. Nature, 515(7528), 582–586. https://doi.org/10.1038/nature13707

Note: This project proposal was written during my Master’s degree for the course Systems Neuroscience. I got a 8.0 (A) for this essay, so I thought I would share.

Background and Aim

Major Depressive Disorder (MDD) is the one of most common neuropsychiatric disorder, with a lifetime prevalence of 14.6% worldwide and 17.9% in the Netherlands¹. MDD is a highly debilitating condition, it has a high economic burden (e.g. around 53 billion dollars in the US alone)² and yet very little is known about its underlying neurobiology – current hypothesis range from inflammation³, mitochondrial dysfunction⁴, HPA axis dysregulation⁵ and gut-brain axis problems⁶. MDD presents many physical effects in the brain, including a reduction in hippocampal and mPFC volume⁷ and increase in activity of default mode network⁸, though the effect size of such human studies is relatively small. Most drugs for MDD have been developed almost 60 years ago, and their efficacy is usually poor for most people – the effect size of pharmacological interventions is 0.35 compared to 0.2 of placebo⁹, many patients are completely treatment-resistant and for the patients for whom the drug works, around 70% present a recurrence of symptoms within 6 years.¹⁰

The lack of effectiveness of traditional pharmacological treatments has led to exploring the development of many alternative treatments, including psychedelic substances. Ayahuasca has been shown to be effective for managing treatment resistant MDD in a randomized clinical trial.¹¹ Although the main ayahuasca active agents, mainly dimethyltryptamine (DMT) and monoamine oxidase inhibitors are well known, the mechanisms through which they act on depression and other affective disorders remain elusive. The effectiveness of ayahuasca and DMT have been also shown in rodents for anxiogenic behavior¹² and non-human primates¹³ for depressive symptoms, suggesting that this drug may be a powerful tool to understand the underlying neurobiology of depression in the mammalian brain.

Marmosets, the model of choice in this study, present many advantages over other non-human primates: they are small, allowing for a simpler housing infrastructure, they have relatively short life cycle (reaching sexual maturity in 18 months and old age around 8 years)¹⁴, their entire genome has been sequenced¹⁵, they can express GCaMP with viral vector injections¹⁶ and, most relevant for this study, they have a well-established model of depression¹⁷, which can be ameliorated with the use of ayahuasca¹³. Marmosets are a great tool for translational neuroscience because of the deep homology in brain circuitry mediating social behavior and reward, extensive use of vision for social signaling and the fact that these primates present depression-like symptoms in nature.¹⁷

The goal of this study is to explore the brain mechanisms underlying: 1) the progression of depressive symptoms and 2) what is the mechanism by which ayahuasca counteracts these symptoms. We will use miniscopes implanted in the vmPFC and field electrodes in the hippocampus – both brain regions previously shown to be associated with depression – and analyze brain activity before social isolation, during social isolation and during treatment (both with an SSRI and with ayahuasca).

Project description

The first thing that needs to be considered for this project is the way in which MDD will be modeled in marmosets. To achieve this, the protocol for social-isolation-induced depression in marmosets, described by Galvão-Coelho et al. (2017)¹⁷ will be followed. This protocol is relatively simmple and has been shown to accurately reflect multiple behaviors and physiological changes observed in depressed primates in the wild as well as in humans. It consists of separating juvenile marmosets from their original homecage, where thay are housed in a family context, and housing them alone in a new homecage (this alone appears to be sufficient to induce depression).

In order to study how activity in the vmPFC and the hippocampus changes throughout the progression of depression in the selected model, as well as evaluating the effect ayahuasca treatment has on these physiological changes and the associated behavioral profile; animals will be divided into four treatment groups, containing 6 animals each (ideally 3 males and 3 females, in order to account for intersexual variation): 1) baseline group, that won’t be exposed to social isolation at any point, 2) a negative control group, socially isolated but receiving only saline solution as treatment, 3) a possitive control group, receiving 0.2 mL/100 g of nortriptyline (an anti-depressant with clinical use) and 4) an ayahuasca group, receiving 6.32mg/kg of ayahuasca (dose adjusted for mass and body surface from a human effective dose.¹¹The animals will be moved into isolation (except for the baseline group) after 4 weeks of daily baseline behavioral and physiological measuremnts, performed after complete recovery from surgical procedures. They will be kept on isolation for 8 weeks and measurements will be obtained daily during weeks 1, 2, 7 and 8. On weeks 9 and 10 drug and vehicle treatments will be applied. Measurements will be collected daily during these two weeks, along with the following two weeks (possible delayed effects). The behavioral parameters that will be meassured are: scent marking, individual piloerection, self grooming, scratching, somnolence, feeding (quantity and frequency) and sucrose solution ingestion. The physiological data we aim to obtain on the other hand includes fecal cortisol levels and changes in brain activity on the vmPFC and anterior hippocampus, as meassured by changes in calcium dynamics and electical activity, respectively.

            For the purpose of meassuring changes in neuron activity in the vmPFC using calcium imaging, AAVs (Thy1S-tTA; TRE3-GCaMP6f) in solution will be injected into Brodmann area 32, a region of the vmPFC associated with depressive behavior in humans and animals, and that has been proven to have depression-relevant connections to the anterior hippocampus in primates¹⁸; this should result in sufficient expression of GCaMP6f and allow us to control said expression through doxycyclin treatment. After 3-4 weeks of incubation time, to ensure stable expression of the introduced constructs a new surgical procedure will be performed, this time to introduce a GRIN lens on area 32 (following the protocol by Kondo et al. (2018)¹⁹, as well as introducing the electrodes on the anterior hippocampus and mounting a headstage²⁰ that will help support both measuring devices.After 5-7 weeks of recovery the base of the miniscope will be fixated to the skull, allowing us to securely mount the device on the animals’ heads. Baseline data collection will begin after a further 4-6 weeks for recovery and habituation. 1h of behavioral measurements and 15min of imaging/electrophysiologydata will be collected every day at the same time,in order to avoid circadian variation. The time of physiological recording is limited to 15min in order to produce managaeable amounts of data and to avoid the effects of photobleaching. Additional to this fecal cortisol levels will be meassured on samples collected daily, in order to obtain information about HPA function.

Image and data analysis

Miniscope

In the miniscope analysis pipeline, it is important to perform an image registration in order to correct for movement artefacts. The NoRMCorre library for MATLAB would be preferred since it is a processing-optimized non-linear motion correction, which means that it takes into account that the brain stretches and compresses as the animal moves around. To do cell segmentation for miniscope data, the commonly used analysis is constrained nonnegative matrix factorization optimized for epifluorescent data (CNMF-E), which creates a mathematical model for the background noise and updates it every frame – reducing false positive and false negatives compare to normal CNMF or other methods such as manual ROI definition or PCA/ICA. [1]^*

Field electrode

Multi-channel data will be processed according to Mohan et al. (2019)²¹. Briefly, initial spike sorting will be done using Mclust 4.3²² and unit clustering will be performed first semi-automatically using klustakwick²³ for initial isolation of units, and then manually, taking into consideration standard parameters such as length of refractory period, spike shape and stability of spiking activity throught the recording. Units can then be clasified based on their spike waveform and analized individually.

Combining miniscope data with electrophysiology data

It is possible to compare miniscope data with electrophysiology data by dimension reduction: essentially, for every cell you capture with the miniscope, the calcium data can be deconvolved and thresholded in order to get unidimensional points of activity. In this way, complex video data would be reduced to the same data time that one would get from single-units in electrophysiology data.[2]^* (Figure 1)

This dataset could be used to compare similar behaviors (when the animal is not moving) over multiple time points (before social isolation, early social isolation, late social isolation, early treatment and late treatment). One could group the data from individual animals (grouping by gender, grouping by treatment type) and assess if there are any commonalities in the development of symptoms of depression and how treatments acts on those patterns of activity. The statistics used will depend on the parameters being assessed – e.g. Mann-Whitney to compare behavioral data (which is not normally distributed), repeated measures ANOVA to assess changes in activity in vmPFC and hippocampus within the same animal, General Linear Models (GLM) Fisher’s post hoc test to assess correlations (activity and behavior).

Whatever changes are found, it would be important to translate the findings both to lower models to ask more mechanistic questions and to humans to potentially come up with new treatments. For example, if it is found that there is a peak of activity in the vmPFC during depression development but it then becomes less active during late stages of social isolation, one could make use of the advanced genetic toolbox for rodents to devise engram-specific/cell-specific studies (e.g. KO of 5-HT2A receptor and assess effectiveness of treatment, fluorescent serotonin probes); but one could also devise a clinical trial in which that brain region is specifically targeted with transcranial magnetic stimulation in patients with familial depression, potentially preventing the progression of symptoms.

It is important to note that the extent of this research project, given the amount of work it requires as well as the time it will take (around 35 weeks, without counting data analysis) goes beyond what could realistically be achieved by a student on a 6 month internship. If necessary, said student could focus only on the social isolation part of the project and make immense contributions to the study.  

Relevance of this project

Although the relevance of rodent models for the study of MDD cannot be denied, it is clear that they suffer from limitations that may be limiting their translatability, both in the context of general neuroscientific research (e.g. phylogenetic distance from humans) and in the specific context of depression, such as most models not reflecting behaviors that naturally occur in wild rodents, a limited social structure²⁴ and considerable divergence in serotonin receptor expression in the brain²⁵. Non-human primate models, such as marmosets, animals with closer phylogenetic relation to humans, relatively complex social structures in the wild²⁴ and display of mood disorder-associated behaviors similar to other larger primates including humans, both in captivity and in their natural environment^17,26 make valuable candidates for future development of models, which could result in a higher translation potential. In this context, beginning to understand the mechanisms that drive depressive behavior in the marmoset models at cellular and systemic levels, as well as the effect potential treatments have on them, could improve our knowledge of the neurological phenomena that drive this disease, as well as to some extent bridging the gap between the knowledge that has been developed in rodents and the current deficiencies in the clinic.

The increased similarity marmosets have with humans, which gives them their great potential value for neuroscientific research such as ours, also makes them a justified source of ethical concern. In order to minimize suffering and maintain an appropriate life quality, marmosets will be housed in comfortable, enriched home cages, with appropriate access to water and food and with an appropriate number of family members (this last one with the exception of when necessary for the experiments), all in compliance with local and international standards (Brazilian Institute of Environment and Renewable Natural Resources, Animal Behavior Society and the International Primatological Society). Humane endpoints consistent with international guidelines will be set for every experiment and consistent monitoring of suffering and well-being will be performed, in order to detect and avoid excessive suffering. In order to increase the power of this study without an indiscriminate increase in sample size, this study uses a paired design, in which the neuronal activity in the same animal is measured before, during and after social isolation and during treatment. This intergroup design setup will provide the most amount of data per animal, allowing for an unraveling of more subtle differences than what it would be possible with an intragroup design.

References

1.        Kessler, R. C. & Bromet, E. J. The Epidemiology of Depression Across Cultures. Annu. Rev. Public Health 34, 119–138 (2013).

2.        Li, X. et al. Depression-Like Behavioral Phenotypes by Social and Social Plus Visual Isolation in the Adult Female Macaca fascicularis. PLoS One 8, (2013).

3.        Felger, J. C. Role of Inflammation in Depression and Treatment Implications. in Handbook of Experimental Pharmacology 250, 255–286 (Springer New York LLC, 2019).

4.        Bansal, Y. & Kuhad, A. Mitochondrial Dysfunction in Depression. Curr. Neuropharmacol. 14, 610–618 (2016).

5.        Keller, J. et al. HPA axis in major depression: Cortisol, clinical symptomatology and genetic variation predict cognition. Mol. Psychiatry 22, 527–536 (2017).

6.        Liang, S., Wu, X., Hu, X., Wang, T. & Jin, F. Recognizing depression from the microbiota–gut–brain axis. International Journal of Molecular Sciences 19, (2018).

7.        Drevets, W. C., Price, J. L. & Furey, M. L. Brain structural and functional abnormalities in mood disorders: Implications for neurocircuitry models of depression. Brain Struct. Funct. 213, 93–118 (2008).

8.        Kaiser, R. H., Andrews-Hanna, J. R., Wager, T. D. & Pizzagalli, D. A. Large-scale network dysfunction in major depressive disorder: A meta-analysis of resting-state functional connectivity. JAMA Psychiatry 72, 603–611 (2015).

9.        Cuijpers, P. et al. The efficacy of psychotherapy and pharmacotherapy in treating depressive and anxiety disorders: A meta-analysis of direct comparisons. World Psychiatry 12, 137–148 (2013).

10.      Verduijn, J. et al. Reconsidering the prognosis of major depressive disorder across diagnostic boundaries: Full recovery is the exception rather than the rule. BMC Med. 15, (2017).

11.      Palhano-Fontes, F. et al. Rapid antidepressant effects of the psychedelic ayahuasca in treatment-resistant depression: A randomized placebo-controlled trial. Psychol. Med. 49, 655–663 (2019).

12.      Cameron, L. P., Benson, C. J., Defelice, B. C., Fiehn, O. & Olson, D. E. Chronic, Intermittent Microdoses of the Psychedelic N, N-Dimethyltryptamine (DMT) Produce Positive Effects on Mood and Anxiety in Rodents. ACS Chem. Neurosci. 10, 3261–3270 (2019).

13.      Da Silva, F. S. et al. Acute effects of ayahuasca in a juvenile non-human primate model of depression. Brazilian J. Psychiatry 41, 280–288 (2019).

14.      Preuss, T. M. Critique of pure marmoset. Brain. Behav. Evol. 93, 92–107 (2019).

15.      Worley, K. C. et al. The common marmoset genome provides insight into primate biology and evolution. Nat. Genet. 46, 850–857 (2014).

16.      Park, J. E. et al. Generation of transgenic marmosets expressing genetically encoded calcium indicators. Sci. Rep. 6, (2016).

17.      Galvão-Coelho, N. L., Galvão, A. C. de M., da Silva, F. S. & de Sousa, M. B. C. Common marmosets: A potential translational animal model of juvenile depression. Front. Psychiatry 8, 1–17 (2017).

18.      Wallis, C. U., Cockcroft, G. J., Cardinal, R. N., Roberts, A. C. & Clarke, H. F. Hippocampal Interaction With Area 25, but not Area 32, Regulates Marmoset Approach–Avoidance Behavior. Cereb. Cortex (2019). doi:10.1093/cercor/bhz015

19.      Kondo, T. et al. Calcium Transient Dynamics of Neural Ensembles in the Primary Motor Cortex of Naturally Behaving Monkeys. Cell Rep. 24, 2191-2195.e4 (2018).

20.      Roy, S. & Wang, X. Wireless multi-channel single unit recording in freely moving and vocalizing primates. J. Neurosci. Methods 203, 28–40 (2012).

21.      Mohan, H. et al. Functional Architecture and Encoding of Tactile Sensorimotor Behavior in Rat Posterior Parietal Cortex. J. Neurosci. 39, 7332–7343 (2019).

22.      Fraley, C. & Raftery, A. E. Enhanced Model-Based Clustering, Density Estimation, and Discriminant Analysis Software: MCLUST. J. Classif. 20, 263–286 (2003).

23.      Harris, K. D., Henze, D. A., Csicsvari, J., Hirase, H. & Buzsáki, G. Accuracy of tetrode spike separation as determined by simultaneous intracellular and extracellular measurements. Journal of Neurophysiology 84, (2000).

24.      Hendrie, C. A. & Pickles, A. R. Depression as an evolutionary adaptation: Implications for the development of preclinical models. Med. Hypotheses 72, 342–347 (2009).

25.      Susser, E., Keyes, K. & Mascayano, F. Healthy pregnancy and prevention of psychosis. World Psychiatry 17, 357–358 (2018).

26.      Barros, M. & Tomaz, C. Non-human primate models for investigating fear and anxiety. Neuroscience and Biobehavioral Reviews 26, 187–201 (2002).

[1]^* In terms of data size, it would not be possible to do whole day recordings, since the data needs to be stored in an SD card and transferred manually to a computer once a week. The timing of 15 minutes once per day would yield enough data to be able to draw inferences about changes in neuronal activity over time and also reduce the amount of GCaMP photobleaching observed. 30 hours of miniscope data would be acquired per week for the 24 animals – since data would not be acquired during the weekends, it would be optimal to have the more processing-demanding part of the analysis running on a powerful computer over the weekend. Another consideration of the miniscope is that we assume it would be possible to record the same neuronal population over multiple days. One main disadvantage of the wireless UCLA miniscope compared to the NiNscope is a lack of focus control in software, so a pilot study would be necessary to assess if daily changes in the focus ring would be necessary.

[2]^* It is important to realize that electrophysiology can provide information about many hundreds of single units, while the miniscope only captures about 100 cells. This sample size difference should be taken into account in the statistical analysis, as well as the delay between neuronal activation and increase in GCaMP fluorescence – which would need to be taken into account in the analysis.

DENTAL ANATOMY AND MARKS OF DEPOSITION

Teeth are comprised of four tissues: Enamel, dentine, cement and pulp. Enamel is the hardest substance present in the human body, consisting mainly (97% w/w*[1]) of calcium hydroapatite. Dentine presents a higher capability of remodelling than enamel, since it contains only about 70% w/w of hydroxyapatite crystals. The pulp contains blood vessels and nerves which supply the teeth and the cementum, with the periodontal ligament, anchors the roots in the sockets of upper and lower jaws.

Because there is no remodelling of enamel after its maturation, which usually occurs after 24 hours after the initial matrix secretion of ameloblasts, its original marks of deposition are retained as cross striations and brown stria of Retzius. Cross striations, also known as short period markers, arise through the circadian rhythm (represented by bands of different levels of mineralization deposited within a single day). The most plausible hypothesis for this pattern of deposition was suggested by Boyde (1964), and further developed more recently by Risnes (1998). They suggested that there would be more carbon dioxide available for incorporation of the enamel during periods of greater metabolic activity, causing enamel to be deposited faster during these moments compared to periods of lesser metabolic activity, which also explains the variation in mineral density found in these markers. The brown stria of Retzius represents a circaseptan cycle, which varies between 7 and 11 days between individuals, but it does not vary throughout the life of a single individual (FitzGerald, 1998). The origin of this cycle is not totally understood. Some authors hypothesise that it is under regulation of melatonin (Hasting and Loudon, 2006), others suggest that it may be a mere stochastic process (FitzGerald, 1995).

Human permanent incisors have about 150 brown stria, canines about 180 and molar contain 120-150 (Bullion, 1987), and the distance between each striae increases towards the incisal or occlusal surface. The wave-like pattern was termed perikymata, subdivided into perikyma grooves and ridges. Counting these markers provides a reliable method of estimating the age of subadults, as tooth eruption follows a strict biological sequence, with a well-known variance between populations. Disruptions of enamel deposition are usually due to genetic enamel defects (collectively referred as amelogenesis imperfecta) or more commonly caused by environmental stress (dietary deficiency, fevers, infections), refered to as hypoplasias.

EVOLUTION IN THE FIELD

Dental morphology plays a key role in multiple aspects of evolution because the dentition is under forceful genetic control. An animal with faulty dentition cannot properly breakdown food in the first stage of digestion, which leads to poorer absorption of energy and ultimately a reduction in the likelihood of its genes being passed onto the next generation.

In their origin, teeth were simple, one-cusped (haplodont), uniform (homodont), conical-shaped structures. Adaptation led to the development of more sophisticated structures, reflecting the versatile dietary aspects of a species. Mammals have heterodont teeth, represented by incisors, canines, premolars and molars: These created occlusal surfaces suitable to a greater variety of food (Phulari, 2013). Herbivorous mammals have molarized premolars, which enhance the surface area and optimize grinding of plant material. Carnivores have dicing/slicing specialized dentition with blade like posterior teeth called carnassials and molars reduced in size or even lost in some species (Hilson, 1986). Omnivores have a more generalized dentition.

Mammals evolved from reptiles in the Mesozoic era, and early mammals presented characteristic triangular-shaped molars (Hillson, 2005). Osborn (1907) hypothesized that the lingual cusp of upper molars and the buccal cusp of lower molars were homologous to the reptilian cone and called these the protocone and the protoconid, respectively. Two other cusps originated as cuspules, and via selective pressure developed into larger structures called the paracone and paraconid at the mesial end of the molar and metacone and metaconid at the distal end in upper and lower molars, respectively. This cusp triad is known as the trigon in upper molars and trigonid in lower molars. In most mammals, the paraconid has been lost in lower molars. An additional ledge of cingulum is present at the distal end of the trigon in primates, known as the talon in upper molars, and the talonid in lower molars. On the upper molars the talon houses a single cusp, the hypocone, but three additional cusps are found on the primate trigonid: hypoconid, entoconid and hypoconulid (Figure 1).

It is now known that the cusp homologous to the reptilian cone in upper molars is the paracone, not the protocone (Butler, 1978), as it was initially suggested by Osborn. Nevertheless, his original nomenclature is still used by most authors, although some prefer to use this terminology and describe the paracone as the primary cusp or cusp 1 (Steele and Bramblett, 1988). Trying to develop a clearer system, Hershkovitz (1978) created an entirely new terminology, replacing the paracone and the protoconid for the eocone and the eoconid. As described by Scott and Turner II (1997), while Hershkovitz has a strong case on paleontological grounds, Osborn’s terminology has such a lengthy history of usage that changing it would create more, rather than less, confusion.

A reduction in the number of teeth and generations of teeth is observed in recent species, possibly as an energetic trade-off for an increase in morphological complexity. Mammals originally presented three incisors in each quadrant of both maxillary and mandibular teeth. The majority of early primates lost one incisor, and by the Eocene there was a reduction in the number of premolars (Swindler, 2004). The modern human dental formula, shared by all catarrhine primates, is 2-1-2-3, while platyrrhines dental formula is either 2-1-3-3 or 2-1-3-2.

How diet influences dental and masticatory morphology is exemplified in early hominins.  Paranthropus presents a more robust mandible, larger temporalis and masseter muscles, and larger posterior teeth with extremely thick enamel when compared to early Homo (Figure 2). This suggests that hard material such as nuts and seeds were an important component of the diet of Paranthropus (Grine and Martin, 1988). Some authors try to relate the trend of reduction of molar size, enamel thickness and masticatory muscles throughout the evolution of the genus Homo to factors such as meat eating (Bunn et al, 2012), the control of fire and cooking (Wrangham, 2009), or even tool use (Teaford, 2007).

EVOLUTION AS A FIELD

Fossilization is an extremely complex and ultimately rare phenomenon, which occurs under very specific circumstances. As a result the fossil record presents very few complete specimens of our distant ancestors. Enamel, the ‘inorganic coating’ of teeth, is the main reason teeth preserve so well in most taphonomic contexts and therefore a high percentage of the fossil record is made up of teeth. Even though other osteological elements are more informative in some respects, they are less likely to be preserved, so palaeontologists and paleoarcheologists have to work with teeth and dental traits in most contexts.

The precise description of dental traits is important, but not easy. Without a standard reference showing variability in the occurrence of dental traits, dichotomized nominal measurements, characterised by absence/presence, were used for many years in the field. In 1920, Ales Hrdlička published the famous ‘Shovel-shaped teeth’, which provided a set of ordinal measurements for the shovelling trait, which included the absence of the trait and three degrees of expression. Other workers had discussed morphological gradations in earlier papers (Black, 1902), but as criticised by Hrdlička, they did not focus on specific trait variance.

Although it is arguable that interval measurements from quantitative or metric data provide precise information with a lesser degree of error, the trend of non-metric observations has prevailed in the field. As described by Scott and Turner II (1997), most dental traits are not easily described by quantitative data, because: ‘(1) they exhibit variation in both form and size, not just size; (2) as three dimensional objects, they often lack the kinds of landmarks that are needed to standardize measurements; and (3) they may show such slight levels of expression, there is no way to measure them in millimetres.’

Basically, dental traits are not metric/non-metric per se, but the absence of precise tools and methods makes the description of most dental characteristics more attainable using qualitative rather than quantitative descriptions. Hanihara (1970) presented a creative way of measuring an ‘essentially qualitative’ trait (hypocone) through quantitative measures.  He photographed of the occlusal surface of the molar, printed an enlarged version and cut the area delimited by the hypocone, which he weighed using a high-precision scale in order to stablish a relative area of the hypocone in relation to the total crown area, which created incredibly robust results. This technique has been since used by many workers (Yamada and Brown, 1988; Bailey et al, 2004; Benazzi et al, 2011).

Since Hrdlička there have been a lot of improvements in this field. Dahlberg (1956) and Hanihara (1961) were the first to develop a dental trait scoring system with plaster plaque references for permanent and deciduous dentition, respectively. This was an incredibly important step in the field: Since an analysis based on scoring non-metric traits is the most feasible approach for characterizing dental traits, a physical plaster model is not only a good way of establishing a visual reference and therefore reduce intraobserver and interobserver inaccuracies, but it is also a powerful learning tool for students in the field.

More recently, Turner II et al. (1991) developed the Arizona State University Dental Anthropology System (ASUDAS), which has been the main scoring system since its original publication. The traits used in this system were chosen because they are easily observable and persistent, meaning they provide relevant information about a population via trait frequency. Many of the ASUDAS traits are invariable throughout hominoid evolution, suggested by Irish and Guatelli-Steinberg (2003), so they can also be used in characterizing non-human primates and early hominins dental morphology (Irish, 2013).

TEETH WEAR AND DENTAL PATHOLOGIES

As discussed before, enamel is not capable of remodelling after its maturation. With time, the forces applied by masticatory muscles on the occlusal/incisal surfaces of teeth cause the enamel to wear down. Patterns of wear can be used to access dietary behaviours. Hunter-gatherers may use their anterior teeth to process hides, causing the incisal surfaces to wear down faster. Early agriculturalists, on the other hand, wore down their posterior dentition faster, due to abrasive particles introduced by processing grains using grinding stones (Smith, 1984). If the process of tooth wear is slow, secondary dentine is produced by remaining odontoblasts in order to protect the pulp.

Another effect of the forces applied on teeth is the slow decrease in crown height, causing a remodelling of the alveolar process (the portion of the jaw which contains the teeth sockets). This results in the continuous eruption of teeth throughout life, eventually causing tooth loss due to lack of bone support (Hillson, 2008). A different cause of tooth loss is periodontitis: Short episodes of inflammation caused by plaque alternated with periods of recovery leading to bone reabsorption and reduction in height of the alveolar process.

Caries is also an important component in tooth loss, but in a different way. Although the inflammation in pulp and periapical tissues caused by caries does not cause the remodelling of bone tissue associated with the loss of teeth, the pain inflicted by it makes human intervention likely to happen in order to remove the carious tooth. Caries is a demineralization process, caused mainly by the accumulation of a specific type of bacteria (Streptoccocus mutans) of the dental plaque (Loesche, 1998). There is a strong correlation between the presence of caries and the increase of sugar consumption during the twentieth century (Thylstrup and Fejerskov, 1994). There is also noticeable decrease in decay rate of children as a result of sugar rationing in Japan and Norway during World War II. The main cause of caries is commonly regarded as sucrose, but the similarity in the molecular structures of simple carbohydrates reflects a small difference in the cariogenicity of fructose, glucose, lactose and sucrose (Hillson, 2008).

The nature of enamel deposition is used in the context of forensic dentistry in various ways.(1) Teeth found in isolation may be matched comparing the brown striae and cross striations to determine the identity of an individual, as the whole dentition of an individual presents, roughly, the same pattern; (2) Historical stressful events (e.g. volcanic eruptions or long periods without food) are able to be matched biologically, if all individuals s dated to a specific period have similar disruption markers. (3) The first brown stria of Retzius, known as Neonatal line and caused by the physiological stress of birth, can be used to determine if infants died  before or after they were born, since prenatal enamel does not usually contain striae (FitzGerald and Rose, 2008). (4) In modern forensics and criminology, in situations where most tissues were destroyed, the extraction of DNA through the pulp and dentin is practical for identification, due to the high chemical and physical resilience of these dental tissues (Datta and Datta, 2012). (5) The mineralized cells present in dental calculus provide an important source of ancient DNA.

With the advance of PCR (polymerase chain reaction), it is feasible to use DNA amounts as miniscule as 1 ng are useful for identification purposes (Sweet, 2001). Dental tartar/calculus retains pathogenic cells, human cheek cells and dietary components such as plant particles, therefore it is useful in many contexts. Because the collection of calculus is a simple and non-destructive procedure, whereas  the most common process of collecting DNA from skeletal remains requires powdering of a portion of teeth or bones, the extraction of DNA from calculus presents a major advantage. Plant particles in tartar are useful for recreating specific aspects of a population’s diet. For instance, starch particles can help understand the methods for cooking and processing food (Henry et al, 2009). A well-preserved dietary record may be used to establish environmental changes, changes in subsistence patterns and migration (Dudgeon and Trump, 2012). It is difficult to extract genetic material from pathogenic agents of ancient skeletal remains, since the amount of DNA and RNA present is very small even in highly infected individuals (Stone, 2008). When it is possible to obtain sufficient pathogenic material from PCR, however, it provides strong evidence for migration of populations and the origin and development of disease throughout human history.

CONCLUSION

The dentition and the characteristics of certain aspects of dental morphology are useful in many scientific fields (embryology, genetics, anthropology, palaeontology, forensic science, et cetera).  If one wishes to acquire more in-depth information about dental anthropology, see Scott and Turner II (1991) and Hillson (1996); Mammalian teeth are explored thoroughly in Hillson (2005), the same for primate dentition in Swindler (2002).

REFERENCES

Bailey, S. E., Pilbrow, V. C., & Wood, B. A. (2004). Interobserver error involved in independent attempts to measure cusp base areas of Pan M1s. Journal of anatomy, 205(4), 323-331.

Benazzi, S., Coquerelle, M., Fiorenza, L., Bookstein, F., Katina, S., & Kullmer, O. (2011). Comparison of dental measurement systems for taxonomic assignment of first molars. American journal of physical anthropology, 144(3), 342-354.

Black, G.V. (1902). Descriptive anatomy of human teeth. White, Philadelphia.

Boyde A. (1964). Structure and development of mammalian enamel. Ph.D. Thesis, Department of Anatomy, London Hospital Medical College.

Bunn H.T, Domínguez-Rodrigo M, Pickering T.R. (2012) How meat made us human: archaeological evidence of the diet and foraging capabilities of early Pleistocene Homo in Eastern Africa. In: Lee-Thorp J, Katzenberg M.A, eds. Oxford Handbook of the Archaeology of Diet. Oxford: Oxford University. In press.

Butler, P.M. Molar cusp nomenclature and homology. In Development, Function and Evolution of Teeth, ed. P. M. Butler and K. A. Joysey. London: Academic Press, pp. 439-454.

Dahlberg A.A. (1956). Materials for the Establishment of Standards for Classification of Tooth Characteristics, Attributes and Techniques in Morphological Studies of the Dentition, Zoller Laboratory of Dental Anthropology, Chicago, University of Chicago..

Datta, P., Datta, S. S. (2012). Role of deoxyribonucleic acid technology in forensic dentistry. Journal of Forensic Dental Sciences, 4(1), 42–46. doi:10.4103/0975-1475.99165

Dudgeon, J.V. and Tromp, M. (2012). Diet, Geography and Drinking Water in Polynesia: Microfossil Research from Archaeological Human Dental Calculus, Rapa Nui (Easter Island). International Journal of Osteoarchaeology. Available at: http://onlinelibrary.wiley.com/doi/10.1002/oa.2249/full [Accessed 16th February 2015]

FitzGerald C. M. (1995). Tooth Crown Formation and the Variation of Enamel Microstructural Growth Markers in Modern Humans. Ph.D. Dissertation University of Cambridge.

FitzGerald C. M. (1998). Do enamel microstructures have regular time dependency? Conclusions from the literature and a large-scale study. Journal of Human Evolution 35:371–386.

Fitzgerald, C. M. and Rose, J. C. (2008). Reading Between the Lines: Dental Development and Subadult Age Assessment Using the Microstructural Growth Markers of Teeth, in Biological Anthropology of the Human Skeleton, Second Edition (eds M. A. Katzenberg and S. R. Saunders), John Wiley & Sons, Inc., Hoboken, NJ, USA. 

Grine, F.E. and L.B. Martin. 1988. Enamel thickness and development in Australopithecus and Paranthropus. In Evolutionary history of the robust australopithecines (ed. F.E. Grine), pp. 3-42. New York, Aldine de Gruyter.

Hanihara K. (1961). Criteria for classification of crown characters of the human deciduous dentition. Journal of the Anthropological Society of Nippon 69, 27-45.

Hanihara, K., Tamada, M. and Tanaka, T. (1970). Quantitative analysis of the hypocone in the human upper molar. Journal of the Anthropological Society of Nippon 78, 200-207.

Hastings M.H., Loudon A.S. (2006). Metabolic regulation: fasting in the dark. Current Biolology 16: R254 –R255.

Henry, A.G., Hudson, H.F. and Piperno, D.R. (2009) ‘Changes in starch grain morphologies from cooking’. Journal of Archaeological Science. 36 (3), 915-922.

Hilson, S. W. (1996). Dental Anthropology. Cambridge: Cambridge University Press.

Hilson, S. W. (2005). Teeth. Cambridge: Cambridge University Press.

Hillson S. W. (2008). Defects of dental development in the enamel of the tooth crown, in Biological Anthropology of the Human Skeleton, Second Edition (eds M. A. Katzenberg and S. R. Saunders), John Wiley & Sons, Inc., Hoboken, NJ, USA. 

Hrdlička, A. (1920). Shovel-shaped teeth. American Journal of Physical Anthropology 3, 429–465. 

Irish J.D., Guatelli-Steinberg D. (2003). Ancient teeth and modern human origins: An expanded comparison of African Plio-Pleistocene and recent world dental samples. Journal of Human Evolution 45, 113–144.

Irish J.D., Guatelli-Steinberg D, Legge S.S., de Ruiter DJ, Berger L.R. (2013). Dental morphology and the phylogenetic “place” of Australopithecus sediba. Science 340(6129), 1233062.

Loesche W.J. (1996). Ch. 99: Microbiology of Dental Decay and Periodontal Disease. In Baron S, et al.Baron’s Medical Microbiology (4th ed.). University of Texas Medical Branch.

Phulari, R. G. S. (2013). Textbook of Dental Anatomy, Physiology and Occlusion. India: JP Brothers Medical Ltd.

Risnes, S. (1998). Growth tracks in dental enamel. Journal of Human Evolution, 35, 331–350.

Scott, G. R. (1991). Dental Anthropology. Encyclopedia of Human Biology 2: 789–804.

Scott, G. R. and C. G. Turner II (1997). The Anthropology of Modern Human Teeth: Dental Morphology and Its Variation in Recent Human Populations. Cambridge: Cambridge University Press.

Swindler, D. R. (2002), Primate Dentition: An Introduction to the Teeth of Non-human Primates, Cambridge: Cambridge University Press.

Smith, B. Holly (1984). Patterns of molar wear in hunter–gatherers and agriculturalists. American Journal of Physical Anthropology 63(1): 39-56. 

Steele, D.G and Bramblett C.A. (1988). The anatomy and biology of the human skeleton. College Station, Texas: Texas A&M University Press.

Stone, A.C. (2008). DNA Analysis of Archaeological Remains, in Biological Anthropology of the Human Skeleton, Second Edition (eds M. A. Katzenberg and S. R. Saunders), John Wiley & Sons, Inc., Hoboken, NJ, USA. 

Sweet D. (2001) Why a dentist for identification? Dental Clinics of North America. 45: 237-51.

Swindler, D. R. (2002). Primate Dentistry: an Introduction to the Teeth of Non-Human Primates. Cambridge: Cambridge University Press.

Teaford, M.F. (2007). What Do We Know and Not Know about Diet and Enamel Structure?, in Evolution of the Human Diet: The Known, the Unknown and the Unknowable. Ungar, P. (Ed.). Oxford, Oxford University Press.

Thylstrup A, Fejerskov O. (1994) Textbook of Clinical Cariology. Copenhagen: Munksgaard.

Turner II. C.G, Nichol C. and Scott G. (1991). Scoring Procedures for Key Morphological Traits of the Permanent Dentition: The Arizona State University Dental Anthropology System. In Kelly M. and Larsen C.S (eds) Advances in Dental Anthropology. Willey-Liss, New York, pp. 13-31.

Wrangham, R. (2009). Catching Fire: How Cooking Made Us Human. Basic Books, 2009.

Yamada, H., & Brown, T. (1988). Contours of maxillary molars studied in Australian Aboriginals. American journal of physical anthropology, 76(3), 399-407.