Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Emergence of abstract rules in the primate brain


Various aspects of human cognition are shaped and enriched by abstract rules, which help to describe, link and classify discrete events and experiences into meaningful concepts. However, where and how these entities emerge in the primate brain and the neuronal mechanisms underlying them remain the subject of extensive research and debate. Evidence from imaging studies in humans and single-neuron recordings in monkeys suggests a pivotal role for the prefrontal cortex in the representation of abstract rules; however, behavioural studies in lesioned monkeys and data from neuropsychological examinations of patients with prefrontal damage indicate substantial functional dissociations and task dependency in the contribution of prefrontal cortical regions to rule-guided behaviour. This Review describes our current understanding of the dynamic emergence of abstract rules in primate cognition, and of the distributed neural network that supports abstract rule formation, maintenance, revision and task-dependent implementation.


The ability of humans and other animals to learn rules enables the classification of relationships and commonalities between events, objects and actions, which can improve perception, learning and decision-making and influence current and future actions. For example, imagine a tourist walking out onto a street in a new neighbourhood or a new country for the first time: learned rules enable this person to efficiently classify and group novel objects (stationary or moving) into behaviourally relevant categories that facilitate opportunities for satisfying their current or future needs for nourishment, transportation and exploration, as well as helping them to avoid potential environmental dangers. In addition, this individual’s social behaviour and interactions will also be shaped and enriched by quickly adapting to learned abstract social rules.

Learned rules are continually refined and updated through experience to become ever more effective in guiding adaptive behaviour. Hierarchies of rules, at different levels of abstraction, shape and structure our perception and enable context-dependent adaptive behaviour in complex volatile environments. Without abstract rules, every piece of information from multiple modalities would need to be analysed and compared with remembered exemplars before the identity, functionality and purpose of a particular object or event could be inferred, imposing undue burden on cognitive resources at the expense of adaptive behaviour1,2,3,4,5,6. Many advanced cognitive abilities in human and non-human primates, such as inferential reasoning, planning, social interaction and flexibility in adapting to a novel situation, are known to be strongly dependent on the formation and implementation of rules5,6,7,8,9,10, and impairments in such cognitive processes have been reported in a range of neuropsychological disorders11,12,13,14,15,16 (Box 1).

Various lines of studies indicate that the emergence of abstract rules for guiding behaviour encompasses dynamic multistage processes, including rule formation, maintenance and revision as well as goal-directed modulation of other cognitive functions. These processes are linked to a diverse range of executive functions2,5 and, accordingly, all involve concerted interaction within distributed neural networks. Psychophysical and neuroimaging studies in humans suggest that the prefrontal cortex (PFC) is heavily involved in the representation and implementation of abstract rules2,17,18,19,20,21,22. Studies in non-human primates have indicated that monkeys also efficiently use abstract rules3,4,5,6,7,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38 to guide behaviour and that neural activity in the primate PFC, as well as other cortical regions, conveys rich information about abstract rules. However, intriguing findings from neuropsychological studies of patients with PFC damage and from lesion-behavioural studies in monkeys suggest that even large lesions in the PFC do not necessarily impair rule-dependent behaviour, and that only specific regions of the PFC might be essential for learning and implementing abstract rules7,39,40,41,42,43,44. The apparently contradictory findings between correlational studies — functional magnetic resonance imaging (fMRI) studies in humans and single-cell recordings in monkeys — and causal (lesion) studies of the PFC have inspired interest in the specific and essential contribution of PFC regions to various aspects of rule-guided behaviour23,43,44.

This Review focuses on the interaction of abstract rules and executive functions2,6 in controlling goal-directed behaviour in primates. We consider studies that have examined the roles of various subregions of the PFC and other posterior brain regions in rule-based behaviour, focusing on evidence from imaging data, electrophysiological studies and other methodologies that supports the existence of comprehensive wide-ranging networks and shared neural mechanisms for abstract rule formation and use in primates. We ultimately propose that a unified framework can be identified that links the formation and implementation of abstract rules with the neural architecture of cognitive control.

Rules: definition and classification

Instances of repetition of spatio-temporal relationships between discrete object features, events and actions present as statistical regularities in the environment. Through repeated experience of these regularities, animals learn to generalize and link these events and objects together, and consequently to extract and apply rules that guide their future responses1,2,10,17,45,46. Rules can of course be classified on the basis of their complexity, but they can alternatively be classified according to the degree of generalizability; hence, we can refer to concrete rules or abstract rules.

Concrete rules describe simple spatio-temporal links between objects, events and actions. Examples of concrete rules include the learned rule that a red light instructs a specific action (such as stopping movement) or indicates a specific outcome (such as a food reward). These links are known as stimulus–response or stimulus–outcome associations. The formation of concrete rules is experience-dependent, meaning that they are typically learnt gradually across multiple instances of experienced positive and/or negative reinforcement (reward or punishment).

Abstract rules are complex and applicable to multiple exemplars (in contradistinction to concrete rules based on stimulus–response associations). Indeed, a distinctive feature of abstract rules is that they are readily generalizable to novel exemplars. Some kinds of abstract rules require some actions to be based on an abstract relationship between objects and actions. For example, the abstract rule that governs whether events are matching versus non-matching is independent of individual encounters and instead describes the scheme in which stimuli can be compared34,47. Abstract rules might also require selective attention to a certain stimulus dimension (such as colour or shape) of an object in particular situations but to alternative stimulus dimensions in other situations. Such overarching abstract rules can be applied to any relevant stimulus — even novel stimuli encountered in an unfamiliar context. Unlike concrete rules, abstract rules describe interactive and causal associations between objects, events and responses. Thus, abstract rules can evoke completely different responses to the same stimulus exemplar depending on the goal and context.

Concrete and multifaceted categories

Categories can also describe associations between objects and their features. Many animals, including mammals48,49,50,51,52, amphibians53, birds9,54 and insects55,56, possess the ability to form and use categories to guide their behaviour. Some categories are defined according to simple boundaries in state space and are manifested through the similarity of perceptual and/or behavioural responses to stimuli categorized as being on the same side of a boundary, but distinct perceptual or behavioural responses to stimuli categorized as being on different sides of a boundary8,20,52. This categorization of environmental stimuli might be innate53,55 or based on experience-dependent learning50,52 and initiated in sensory association areas wherein relevant sensory features are processed and integrated into perceptual and mnemonic functions20,50,57,58,59,60.

Concrete categories are bound to individual exemplars or individual features in the definitions of category boundaries, often within a single sensory modality. However, more complex, multifaceted categories (such as animate versus inanimate objects) also exist. Multifaceted categories are not bound to individual exemplars or features; instead, they involve boundaries based on integrated features of stimuli that are often perceived using different modalities, which might include sensory information (colour, shape or sound), functionality and interactions with the environment52,58. Accordingly, multifaceted categories readily facilitate generalization to novel encounters; they offer increased subtlety in terms of evolving category boundaries as experience accrues, and they offer increased flexibility for subsequent redefinition of category boundaries across stimulus sets in the face of markedly changed experiences and/or outcomes. Multifaceted categories are conceptually and functionally similar to abstract rules in that they are capable of guiding complex goal-directed behaviour; moreover, abstract rule-guided behaviour can act upon categorized sensory information. The prefrontal neurocircuitry (discussed below) conveys rich information about abstract rules, multifaceted categories and their interactions.

Abstract rules in primates

Adaptive behaviour in humans is dependent on multiple cognitive processes that flexibly draw upon abstract rules for both learning and guiding goal-directed behaviour1,5,6. Although the presence of comparable cognitive abilities in non-human animals has been debated61, studies conducted in non-human primates indicate clear behavioural manifestations of abstract rules5,6,7,26,27,28,33,37,38,39,40,49,61,62. Abstract rules represent an evolutionary adaptation that promotes survival and reproduction and improves interactions with the environment in many species; however, abstract rules are particularly well developed in primates5,7,10,20. Indeed, non-human primates show a high level of flexibility and generalizability in acquiring and implementing abstract rules such as object matching versus non-matching34,35, colour matching versus shape matching5,6,7,26,27,28,33 and matching to number37,38 to guide their behaviour.

Psychophysical, neuroimaging and electrophysiological studies in humans and monkeys have addressed the neural substrates and mechanisms that underpin the formation and implementation of abstract rules. In the following sections, we discuss relevant evidence from neural activity recordings and brain lesion studies in the context of some of the most routinely used experimental tasks.

Evidence from neuronal recordings

Matching and non-matching rules

Delayed matching to sample (DMS) and delayed non-matching to sample (DNMS) tasks require rule-based comparisons of the sameness of or difference between stimuli that can be generalized to multiple exemplars, including novel items. To investigate the neuronal representation of this kind of abstract rule-guided behaviour, macaque monkeys were trained to apply DMS and DNMS rules to visual items34,35,63 with a familiar cue at the start of each trial indicating the relevant rule to be applied (Fig. 1). Monkeys needed to be able to maintain the rule information across a delay period to be able to apply the rule when the sample and test items were shown. Across studies, single-neuron activity was recorded in the dorsolateral prefrontal cortex (DLPFC), ventrolateral prefrontal cortex (VLPFC), premotor cortex (PMC), inferior temporal cortex (ITC), dorsal striatum and orbitofrontal cortex (OFC) while monkeys performed the task. During the delay period, rule-dependent neural activity was observed in all of the regions listed34,35,63 (Fig. 1). However, rule-dependent neural activity was strongest in the PMC and weakest in the ITC.

Fig. 1: Neuronal representation of abstract rules and multifaceted categories.

a | Events in each trial for a task requiring application of an abstract rule (matching versus non-matching). The cue instructing the relevant rule in each trial was selected from two different modalities to control for cue-specific neural responses. To obtain a liquid reward, monkeys were trained to release a lever if either the sample and test objects were identical (match rule) or if the objects were different (non-match rule)34. Therefore, the monkeys needed to remember both the sample and the relevant rule (match versus non-match) during the delay period. b | The activity of a single neuron in the monkey dorsolateral prefrontal cortex. The mean activity in two matching conditions (blue and green lines) and in two non-matching conditions (orange and red lines) is shown. Neuronal activity during the delay period was dependent on the abstract rule, irrespective of the object type or the cue used to instruct the rule. c | In a similar experiment that applied a matching rule, category boundaries were defined by dividing 360° of visual motion directions into two 180° categories separated by a learned boundary (dashed line). The six coloured arrows indicate motion directions belonging to category 1 (blue) and category 2 (green). Trials in the delayed match to category task begin with a gaze fixation period followed by presentation of a sample motion stimulus. After a delay, the monkey is presented with a test motion stimulus. To receive a reward, monkeys must indicate (by releasing a manual touch bar) whether the test stimulus is a categorical match to the sample stimulus. If the first test is a non-match, it is followed by a brief delay and a second test stimulus that is always a match to the sample (and so requires a response)8,50. d | Recordings from two individual category-selective neurons in the lateral intraparietal sulcus cortex (LIP) and prefrontal cortex (PFC). Traces indicate the neurons’ firing rates, averaged across trials, in response to each of the six motion directions. Blue and green traces indicate firing rates in response to motion directions in categories 1 and 2, respectively, during the task. Note the similarity of firing rates in response to stimuli in the same category, which are distinct from firing rates in response to stimuli in different categories8,50. RF, neuronal response field. Parts a and b adapted from ref.34, Springer Nature Limited.

Human fMRI studies that compared brain activity during a task requiring a similar abstract rule with activity during a task using a concrete rule17,64 confirmed that rule-dependent activation occurred in a widespread network involving the left posterior VLPFC, pre-supplementary motor area, left frontopolar cortex and left posterior parietal cortex (PPC). Although single-neuron recordings in monkeys found abstract rule encoding in both the DLPFC and the VLPFC, human studies have found no evidence of rule-related activation in the DLPFC17,64. Another human fMRI study that contrasted DMS and DNMS65 tasks found considerable neural activity in the left PMC, bilateral OFC, left dorsomedial thalamus and left precuneus. Prefrontal-projecting cerebellar lobules also show abstract rule-related activation66.

Number rules

The number and numerals are abstract concepts that describe a quantity of items independently of their identity and component features and can be generalized to novel exemplars. In a series of studies, macaque monkeys have been trained to compare the number of items on a display, irrespective of the physical appearance of the items, and to deliver a response if they matched in number21,36,37,38,67,68,69,70. One study that used this approach showed that more than 30% of recorded cells in the DLPFC and VLPFC encoded information about the number, generalized across changes in the physical appearance of the displayed items36,38. Individual neurons exhibited peak activity for one specific number and declined in firing rate as the quantity of items diverged from this preferred number36,38. Follow-up studies indicated that a substantial proportion of neurons in the intraparietal sulcus (IPS) of the parietal cortex encode numerical information21,36,37,38,69. Interestingly, DLPFC neurons and IPS neurons represented numerical information conveyed by both auditory and visual stimuli, which suggests that these regions encode multimodal representations of abstract numerical information71.

In another study67, monkeys were trained to compare the sizes of a sample set and a subsequently presented test set of objects, and to respond in accordance with either a ‘greater than’ or a ‘less than’ abstract rule, depending on which rule was instructed by sensory cues. Single neurons in both the DLPFC and the dorsal PMC encoded the abstract rule, although rule selectivity was enhanced in dorsal PMC neurons72. Single-neuron activity has also been recorded in the medial temporal lobe of neurosurgical patients performing a numerical task70, and revealed that individual medial temporal lobe neurons encoded either non-symbolic (dot array) or symbolic (numerals) number concepts, but not both. Human imaging studies have also shown activation in a frontoparietal network during the performance of numerical judgements36,37. Bilateral IPS and dorsal and ventral parts of the lateral PFC were among the most commonly activated brain regions during numerical cognition37,73.

These studies suggest that homologous brain regions (mainly the lateral PFC, PPC and PMC) in monkeys and humans are involved in encoding two different abstract rules (number and match versus non-match)21,36,37,38,46,67,68,69,70,73. A set of prefrontal neurons in monkeys were shown to respond to training with a set of multifaceted categories by encoding those categories. However, after training with new sets of categories, the same neurons altered their response tuning to encode the most recently reinforced (trained) categories instead74. This finding indicates a high level of experience-dependent plasticity of category representation in the prefrontal neurocircuitry. The process of learning abstract rules (and presumably of prefrontal neuron encoding) takes longer in monkeys than in humans. The stronger rule-encoding response observed in dorsal PMC neurons might result from extensive task training, which results in rule encoding becoming more intimately integrated with regions involved in generating motor responses63. Therefore, the PFC and parietal cortex might be principally involved in the extraction and consolidation of abstract rules, whereas the implementation of existing abstract rules might be largely handed over to other regions such as the PMC. This hypothesis could be investigated in future studies by comparing neuronal populations involved in early versus late training. The consequences of lesions in the PMC on the formation and implementation of these abstract rules also remain unclear.

Flexible shifting between rules

In the match versus non-match tasks already discussed (Fig. 1), abstract rules were instructed by a cue and no activity was detected in the DLPFC (superior–lateral prefrontal areas)64. However, abstract rule tasks that have elicited DLPFC activity in humans include those in which abstract rules are inferred rather than instructed75 or in which abstract rules are freely chosen76. In a rare human electrophysiological study77, single-neuron activity in the PFC was recorded while neurosurgical patients performed a task involving a choice between which of two abstract rules was implemented for matching visual objects (functionality matching or similarity matching). In this study, single neurons in the DLPFC encoded the abstract rule independently of the identity of the visual object or the manual response.

Perhaps the most thoroughly established paradigm that involves non-cued selection of abstract rules and flexible shifting between such rules is the Wisconsin Card Sorting Test (WCST), which is routinely used in clinical neuropsychological assessments to assess cognitive flexibility in shifting between abstract rules (colour, shape and number matching). In the WCST, the relevant rule for matching is changed without explicit cueing when correct performance under the current rule meets a predetermined criterion, after which test participants need to discover the new rule by trial and error28,31,33. Importantly, macaque monkeys can also perform a computerized close analogue of this task3,5,6,7,26,27,28,29,31,32,33,78. Monkeys show impressive capabilities in shifting between rules and in generalizing rules to new exemplars, which confirms that they can truly apply abstract rules31,33. Detailed neurophysiological investigation has revealed that substantial proportions of macaque DLPFC neurons encoded the abstract rule within and between trials31. Moreover, the magnitude of rule-related neuronal activity was also related to overall behavioural efficiency in adapting to rule shifts31. In another set of studies that used a different analogue of the WCST, the activity of single neurons in the PPC and posterior part of the VLPFC (inferior arcuate sulcus) represented rule shifts79,80 (Fig. 2a,b). Temporary inactivation of the VLPFC region by muscimol (a GABA receptor agonist) impaired rule-shifting, suggesting an essential role for this region in supporting cognitive shifting between rules80. In macaques, a study using a third analogue of the WCST (all WCST analogues used in monkeys are structurally similar to the human WCST and are presumed to recruit the same cognitive functions) showed that single neurons in the OFC, dorsal striatum and ventral striatum encode behaviour-guiding rules81. The OFC neurons also encoded activity relating to rule-shifting (switch-to and switch-away signals) during this task. Shift-related activity was observed in both ventral striatum and dorsal striatum neurons, albeit later in the trial-and-error period in the dorsal striatum than in the ventral striatum. These observations suggest a functional handoff from the ventral striatum to the dorsal striatum during rule-reconfiguration processes and again demonstrate the involvement of neuronal circuitry beyond the PFC81,82,83.

Fig. 2: Neural representations of shifts between abstract rules.

a | Events in each trial of a computerized analogue of the Wisconsin Card Sorting Test (WCST)79,80, in which a sample stimulus was shown at the centre and three test items appeared on the left, right and above the sample. The monkey was trained to select (by making a saccadic eye movement towards) one of the test items that matched the sample in either colour or shape. In non-shift trials, the monkeys needed to continue to apply a previous rule to receive a liquid reward, whereas in shift trials they needed to abandon the previous rule and apply an alternative rule. When the matching rule changed without any notification, the animals committed an inevitable error because they applied the previously relevant rule. These inevitable-error trials were not rewarded; monkeys had to apply the alternative matching rule in the following ‘shift trial’ to receive a reward. b | Activity of a single neuron in the posterior parietal cortex encoding the shift between abstract rules while monkeys performed an analogue of the WCST that required shifting between two rules (matching based on either colour or shape)79,80. Activity of this neuron was higher in shift trials than in non-shift trials. c,d | Functional magnetic resonance imaging of monkeys (part c) and humans (part d) performing an analogue of the WCST84,87. Schematic diagrams show activated areas of the ventrolateral prefrontal cortex in monkeys and in humans, contrasted between shift and non-shift trials. Green and blue arrows show the principal sulcus and inferior arcuate sulcus, respectively, in the monkey. The corresponding cytoarchitectural regions are also shown for humans. Parts a and b adapted with permission from ref.79, Elsevier. Parts c and d adapted with permission from ref.84, AAAS. Parts c and d cytoarchitectural maps adapted with permission from ref.85, Elsevier.

Neuroimaging studies provide complementary evidence for the involvement of distributed networks in the WCST and other tasks that require shifting between abstract rules. In an fMRI study84, monkeys and humans performed a computerized analogue of the WCST (Fig. 2c,d). In monkeys, shift-related activity was observed bilaterally in the anterior bank of the descending ramus of the arcuate sulcus, left IPS, bilateral posterior cingulate cortex, precuneus and insula (Fig. 2c). In humans, shift-related activation was observed bilaterally in the posterior part of the inferior frontal sulcus (mainly areas 44 and 45)85, the inferior parietal lobule and the anterior insula84,86,87 (Fig. 2d). Other human imaging studies conducted in the context of the WCST and other rule-shifting tasks have shown shift-related activation in a distributed neural network including the VLPFC, PPC, DLPFC, anterior cingulate cortex (ACC), caudate nucleus and anterior insular cortex88,89,90,91,92,93,94,95. Activation related to cognitive rule-shifting was observed mainly in the left hemisphere87,88,96.

Abstract response strategies

Achieving goals in volatile environments demands dynamic selection between and implementation of different abstract rules and strategies. Examples of abstract response strategies are repeat–stay and change–shift (which describe making the same or a different response to successive stimuli, respectively); adoption of such strategies can improve animals’ task performance beyond that mediated by stimulus–response association learning. Neurons in the DLPFC and VLPFC encode abstract response strategies towards visual stimuli97 and lesions in the VLPFC and OFC (which are the PFC regions that receive input related to visual objects from the ITC) impair performance based on abstract response strategies98. Abstract response strategies share many features with abstract rules; for example, both involve generalization of responses to exemplars and their application to novel stimuli.

Other approaches to studying the interplay between abstract rules and response strategies in monkeys have involved training animals to achieve a goal or obtain a reward by choosing to use one of several possible tasks. Neurons in the DLPFC, VLPFC and PPC (mainly in the IPS) of trained animals were shown to encode the task (or rule)99,100,101,102,103. Rule-dependent neuronal activities emerged earliest and were strongest in the PFC, suggesting that rule-related information was selectively transmitted in a top-down direction from PFC to PPC neurons, indicating communication between the nodes of an extended network72,104,105. Moreover, temporary inactivation of VLPFC neurons by muscimol impaired the monkeys’ performance25.

Task-related bilateral activation in the PFC and parietal cortex has also been shown in human imaging studies88,89,91,92,93. Task-switching in humans per se is beyond the scope of this Review, but we draw attention to the fact that different bilateral inter-regional interactions between frontal pole and posterior frontal regions reflected the participants’ preparation for applying distinct abstract rules or tasks during the delay period within a working memory task106. This observation suggests that different network interactions underlie adaptive switching between rules, strategies and tasks in humans.

Interactions of rules and categories

Complex behaviour entails context-dependent interaction of abstract rules and categories when a particular rule is applied to multifaceted categories. For example, a matching or non-matching rule might be applied according to the category definition boundaries52,58 but not based on an individual exemplar. In real life, multiple categories and rules are dynamically retrieved and interfaced to coherently optimize perceptions and actions in complex multimodal episodes. For example, many objects with different features might be classified within the category of ‘knife’. Visual feature category boundaries for ‘knife’ might be stored in the ITC simultaneously with semantic category boundaries in the perirhinal cortex. However, some objects classified as knives might further be categorized as dangerous (in the hands of a threatening individual), appetitive (if a famous chef is using it to prepare food) or worrying (if an infant has picked it up in error). Therefore, interactions between the behavioural context and abstract rules such as matching or non-matching might dictate completely different responses to categorized items.

In a series of studies, monkeys were trained to categorize morphed visual stimuli into arbitrary multifaceted categories such as dogs and cats46,51,52,58,74. Monkeys efficiently learned to apply a matching to category rule based on visual categories and activity was recorded across different studies in DLPFC, PPC, ITC and lateral PMC neurons. A substantial proportion of neurons in these brain regions encoded the category information by showing large activity differences between categories but similar responses to stimuli within each category46,51,52,58. Moreover, the monkeys were able to learn new category boundaries within the same stimulus space, indicating that categorization was both dynamic and flexible52. In another series of investigations, neuronal activity was recorded while monkeys applied matching rules based on two different categories of motion direction with regard to moving dots on a display (Fig. 1); neurons in the IPS encoded the category information, whereas neurons in middle temporal areas showed direction selectivity. These findings suggest that information about motion direction is conveyed through middle temporal areas to the PPC, where the category representation might be encoded8,50. Further studies using this motion categorization task revealed that neuronal category selectivity emerged with a shorter latency in the LIP, and that category selectivity was stronger in the LIP than in the PFC107 (Fig. 1). These studies support the hypothesis that a distributed network involving PFC and PPC regions supports dynamic interactions between abstract rules and multifaceted categories.

Insight from neuronal recordings

Cross-species studies indicate great similarity in rule representation and shift-related activation within cytoarchitecturally homologous regions in monkeys and humans (Fig. 2c,d). Different aspects of rule-based behaviour, such as preparation for an upcoming task31,102,106, working memory of the relevant rule31,32,34,38,63,69,72,86,87,88 (Figs 1, 3a), rule-shifting79,80,84,86,87,88 (Fig. 2), responses to feedback31,92 (Fig. 3b) and rule-based action selection31,78, differentially activate distributed cortical and subcortical regions89,108 (Fig. 4).

Fig. 3: Prefrontal cortical neurons represent interactions of abstract rules and executive functions.

a | Monkeys performed a computerized analogue of the Wisconsin Card Sorting Test (WCST) in which they had to match a randomly selected sample to one of three test items either by colour or by shape, and point to the correct item on a touch screen. A liquid reward was provided for correct responses and an error signal was shown for erroneous responses. The relevant matching rule remained consistent within a block of trials but changed without notice when correct responses reached a predetermined 85% threshold. The relevant rule and its frequent changes were not cued; therefore, monkeys had to identify the rule shift by interpreting the feedback in the context of the applied rule. In each trial occurring after an inter-trial interval (ITI), a start cue appeared, and if monkeys pressed a bar, the start cue changed to a small fixation point. Presentation of the fixation point (at 700 ms) was followed by appearance of a sample image (at 630 ms). Then, three test items appeared beside and below the sample image and the monkeys touched the test item that matched the sample in either colour or shape. b | Peri-stimulus histogram showing the activity of a single neuron in the dorsolateral prefrontal cortex (DLPFC) while monkeys performed the WCST. Neuronal activity is aligned at onset of the start cue. Activity of this neuron during the ITI was modulated by the abstract rule. As this study did not include a cue to indicate the relevant rule, this neuronal activity might represent working memory of the abstract rule. c | Activity of another DLPFC neuron shown during a similar test including feedback. After monkeys committed perseverative errors (matching based on the irrelevant rule), a visual error signal was shown for 500 ms. Neuronal activity is aligned at onset of the error signal. Although the same error signal was used in relation to both colour and shape trial blocks, the neural response to the error signal was modulated by the abstract rule. Parts b and c adapted with permission from ref.31, Society for Neuroscience.

Fig. 4: Rule-guided behaviour in primates.

Prefrontal and medial frontal regions of human and macaque monkey brains (not to scale), with colour-coded cytoarchitectural divisions shown in lateral, orbital (ventral) and medial views. Human neuroimaging studies17,64,65,75,84,86,87,104,106,120,121,144 as well as neuropsychological and/or lesion studies4,5,7,27,98,109,110,111,112,113 and single-neuron recordings29,31,34,63,78 in non-human primates have revealed dissociations between key regions in distributed frontal networks that underlie rule-guided behaviour. ACC, anterior cingulate cortex; CC, cingulate cortex; DLPFC, dorsolateral prefrontal cortex; DMS, delayed matching to sample; DMNS, delayed non-matching to sample; OFC, orbitofrontal cortex; OLF, olfactory bulb; PFC, prefrontal cortex; PMC, premotor cortex; VLPFC, ventrolateral prefrontal cortex; WCST, Wisconsin Card Sorting Test. Cytoarchitectural maps adapted from ref.85, Elsevier.

These observations suggest that overarching abstract rules influence various cognitive functions through different neural circuits that support rule-guided behaviour. The consistent activation of these regions by different types of abstract rules refutes the concept that regional specializations represent particular types of abstract rules, and instead suggests a shared neural substrate. However, as we discuss later in this Review, the formation and implementation of abstract rules are closely related to executive functions and therefore common activation of these brain regions might reflect the involvement of such functions109. Accumulated evidence indicates that abstract rule encoding is not restricted to a particular prefrontal region, but rather is represented across widespread distributed neural networks, including different compartments of the PFC, PPC and, particularly, regions around the LIP, PMC, anterior insular cortex and anterior striatum. Moreover, the network is not static and can be dynamically reconfigured to reflect evolving goals, new rules and the experience and expertise of the individual with the underlying rules.

Evidence from interventional studies

Neuronal activity recordings and fMRI studies provide only correlational evidence and cannot confirm whether a given brain region has an indispensable role in a particular cognitive function; for that, interventional studies are required, such as brain lesions, reversible inactivation or microstimulation (which can be electrical or optogenetic). In concert with behavioural studies in humans and in animal models, interventional studies have the potential to evaluate the causality of regional contributions to cognition and behaviour1,7,26,28,29,59,110. These studies have provided evidence for a crucial role of the PFC in the formation and application of abstract rules.

Concrete rules

In monkeys, lesions in the OFC, VLPFC and DLPFC, particularly in the periarcuate region, result in impaired learning of concrete rules, such as stimulus–response associations98,111,112,113,114. Neuropsychological examinations of human patients has also shown that large lesions in the PFC can result in deficits in the acquisition of stimulus–response associations115,116. These findings indicate that in these two primate species, the PFC might be necessary for learning concrete rules that require the association of a visual or auditory stimulus with a particular action.

Cued abstract rules

In monkeys, lesions in the VLPFC and OFC, but not in the DLPFC, result in impaired performance in DMS and DNMS tasks in which the abstract rule is not selected by the animal117,118,119,120. These findings are consistent with human neuroimaging studies of cued, rule-guided behaviour (DMS versus DNMS) showing activations in the OFC and VLPFC but not in the DLPFC17,64,121. Other studies in monkeys showed that large lesions in the DLPFC impaired rule acquisition and task performance in DNMS trials122, and showed that inactivation by cooling targeted to the DLPFC impairs DMS task performance, particularly when long delay intervals are used120. The DLPFC findings are consistent with neuronal activity recordings showing that DLPFC neurons convey information about the objects or locations during delay periods2,34,123 and might suggest increased involvement of the DLPFC when DMS or DNMS tasks demand short-term (working) memory32,34,120,122. Lesions or inactivations that are more circumscribed are required to resolve the specific contributions of different prefrontal regions.

Uncued abstract rules

Monkeys with lesions in the DLPFC, ACC or OFC show marked impairment in performing WCST analogues, which require uncued abstract rules to be inferred from assessments of behavioural outcomes7. Deficits in performing the WCST and other rule-shifting tasks have also been reported in patients with brain lesions, particularly when the lesion includes the DLPFC, VLPFC, OFC or ACC6,39,40,41,42,124,125. However, the WCST is a multifaceted test in which the capacity to flexibly shift between rules depends on the interplay of multiple cognitive functions, such as maintaining rule information in working memory, inhibition of irrelevant rules and assessment of behavioural outcomes2,5,6,40. Therefore, impairment in any one of these functions might underlie the observed performance deficits in individuals with extensive brain lesions. Indeed, lesion studies in monkeys have shown that not all PFC lesions result in impaired rule-dependent behaviour7,26,27. Specific subregions of the PFC, when lesioned, have dissociable effects on adaptive rule implementation7,26,29 (Fig. 4). This observation might explain why findings from neuropsychological examinations of patients with brain lesions have been largely inconsistent in specifying which subregions of the PFC contribute to different aspects of rule-shifting ability in the WCST. One early study in humans emphasized the crucial role of the DLPFC in WCST performance40, but follow-up studies have variously indicated that only patients with lesions in the ACC39, OFC41 or DLPFC42 show impaired WCST performance. The findings of a large-scale, comprehensive study that examined patients’ cognitive performance across multiple tests (including the WCST) of executive functions indicated that only lesions in the medial wall of the PFC resulted in impaired test performance126.

Linking cognitive deficits observed in neuropsychological assessments with particular brain regions is not straightforward in human studies because of inconsistencies in lesion extent across examined patients, the absence of pre-lesion test performance data and variability in the elapsed time after lesion onset. Therefore, animal models have enabled the effects of circumscribed brain lesions to be linked with behavioural deficits under tightly controlled experimental conditions7,26,28,29,43,44,78,127. In a series of such studies, monkeys were trained to perform a computerized WCST analogue. Lesions within the DLPFC, ACC and OFC, but not in the posterior cingulate cortex, frontopolar cortex or superior DLPFC, resulted in impaired cognitive flexibility in shifting between abstract rules7,27. Detailed behavioural assessments indicated that animals with DLPFC lesions could not retain the correct rule in working memory in modified WCST trials in which long inter-trial intervals were imposed. The effects of other lesion sites on WCST performance were dissociable: OFC lesions specifically affected the assessment of behavioural outcome and rapid relearning of rule value, whereas ACC lesions influenced the selection of rule-based actions3,4,6,7,26,27,28,29,78. Notably, lesions in the DLPFC, ACC or OFC did not impair rule-based matching in control tasks when the rule remained consistent within a daily session (that is, in a stable environment)7,28.

Rule-shifting in macaques performing a WCST analogue can be considered to involve an extra-dimensional shift, because attention must shift from one dimension of an object to another (such as from colour, required for colour matching, to shape, required for shape matching) and vice versa. Studies that have contrasted extra-dimensional and intra-dimensional shift modalities have shown that lateral PFC lesions in marmosets result in impairment of extra-dimensional but not intra-dimensional shifts or stimulus–reward reversal learning; however, reversal learning is impaired after OFC lesions24. The discrepant findings of these studies between species (namely that shifting between rules is impaired in macaques with OFC lesions whereas extra-dimensional shifts are unimpaired in marmosets with OFC lesions) might be explained by task differences. In both intra-dimensional and extra-dimensional shifts, learning is based on multiple repetitions and associations of the same stimuli with the reward; however, in the WCST, abstract rule values are rapidly and frequently updated and therefore are dependent on the OFC. A separate study using another WCST analogue also showed that large lesions including the DLPFC and/or specifically the superior DLPFC impair the ability of macaque monkeys to shift between rules122.

Multifaceted categories

The effects of large lesions in the lateral PFC on responses based on multifaceted categories have also been examined in monkeys127. An exemplar from one of two different multifaceted categories of visual stimuli (cats or dogs) was presented as a large centrally presented background image while monkeys manually responded to a change in the colour of a centrally presented small cue superimposed over the large background image to obtain a liquid reward. Low-incentive category stimuli (such as cars) indicated small and delayed rewards, whereas high-incentive category stimuli (such as trucks) indicated large and immediate rewards. Although successful completion of the task was not dependent on the information conveyed by the background category exemplars presented, the category information still influenced the monkeys’ behaviour, in that more errors and longer response times were apparent when small and delayed rewards were expected. Large lesions in dorsal and ventral aspects of the lateral PFC left the influence of the category cues on performance intact, and monkeys also learned new categories as efficiently as they did before lesioning44,127. This intact ability to perform tasks modulated by multifaceted categories seems to contradict other findings showing rich representation of multifaceted categories in the activity of PFC neurons52. The intact ability might have been supported by association areas involved in the processing of visual features128. Moreover, object-level representations of exemplars are believed to be sited in the perirhinal cortex129. Given that the perirhinal cortex projects more heavily into the OFC than into the VLPFC (which receives increased projections from the inferotemporal cortex, a site of feature-based representations), task performance in the studies involving reward association and anticipation44,127 might mainly depend on processing in the intact OFC, whereas task performance in studies where matching based on category information is used for action selection might mainly depend on processing in the VLPFC52. According to this view, either DLPFC processing or VLPFC processing might predominate when interactions between rules and multifaceted categories are necessary for action selection and achieving a goal.

Insights from lesion studies

Lesions in cytoarchitecturally homologous regions in monkeys and humans lead to comparable deficits in rule-guided behaviour. In addition, both neuronal activity recordings and lesion studies indicate that the ability to shift between rules depends on these cytoarchitecturally homologous regions, particularly the posterior part of the inferior frontal sulcus in humans and the inferior prefrontal convexity in monkeys84,87. Imaging studies have commonly indicated activation of the insular cortex in both humans and monkeys during the WCST and other rule-shifting tasks84,87. However, limited knowledge is available about rule-related neuronal activity in the insula or the consequences of selective bilateral insular lesions on rule-guided behaviour.

Lesions in PFC regions impair forms of adaptive behaviour that involve dynamic, context-dependent shifting between abstract rules, but do not impair the consistent implementation of abstract rules in unchanging and steady environments (when no rule shifts are required)7,26. This difference emphasizes the crucial role of the PFC in flexible rule-guided behaviour in changing environments (Fig. 5). Although large PFC lesions might result in impaired overall performance of rule-guided behaviour, localized lesions in frontal regions such as the DLPFC, VLPFC, frontopolar cortex, ACC or OFC have revealed dissociable roles of these regions in rule-guided adaptive behaviour (Fig. 4). Specifically, the DLPFC is implicated in the mnemonic aspects of rule-guided behaviour, such as working memory of a rule or the memory of crucial events such as previously committed errors7,28,32,40,125; the OFC is implicated in the assessment of decision outcomes (reward, punishment and error) and, consequently, in updating the validity of rules7,29,41; the ACC is implicated in rule-guided action selection and in monitoring the demand for cognitive control3,4,7,39,78; and the VLPFC is involved in rule-shifting (that is, adopting new relevant rules and abandoning irrelevant rules)7,84,87,88,96. These interactive executive functions are coherently and dynamically recruited and guided by overarching abstract rules to support adaptive behaviour in volatile environments. In this proposed neural architecture of rule-guided behaviour, abstract rules continually interact with a wide range of executive functions and these processes are mediated through distributed networks that include various PFC regions as key hubs5,6,7,27,29,41,42 and extend to other regions, such as the PPC and insular cortex.

Fig. 5: Abstract rules emerge from dynamic multistage processes.

Successive stages of rule formation and implementation are shown along with the possible contributions of the prefrontal cortex (PFC) in primates. Rule-guided behaviour is challenged in new or rapidly changing environments, and therefore requires integration with executive functions and involvement of the PFC.

Stages of abstract rule development

Various lines of studies in humans and monkeys suggest that different PFC regions have dissociable roles in supporting rule-based behaviour (Fig. 4). Here, we propose a unified framework to describe the neural architecture of rule-based control of primate behaviour (Fig. 5). The PFC contributes to multiple stages in the emergence of abstract rules in primates, which are considered in turn below. The involvement of executive functions and contributions of the PFC to rule-guided behaviour in stable and changing environments are also addressed.

Rule formation

Formation of concrete rules is dependent on perceptual and mnemonic processes in association cortices20,32,128 and is mainly guided by reinforcement learning, mediated by subcortical structures such as the striatum47,57,81,82,130 and hippocampus95,131,132,133,134. The PFC might also be necessary in the early stages of rule extraction to establish associations between objects, events and actions, particularly when stimuli need to be associated with an action to achieve a particular goal111,112,113,114,116 (Fig. 5). Psychophysical and imaging studies indicate that abstract rules might become established by linking together different concrete rules and integrating them with instructions and information for achieving a particular goal49,135,136. The processes involved in abstract rule formation are goal-oriented and cognitively demanding. Compared with simple association learning, abstract rule learning entails a longer time span, requires allocation of executive control, involves more elements (that is, objects, events and their interactions) and might necessitate multimodal processing and extensive retrieval of memory as well as the integration of these processes (Fig. 5).

The PFC is necessary for learning new abstract rules, particularly when achievement of specific goals requires exploration of the contingencies and values (that is, costs and benefits) associated with events in the environment2,137. For example, imaging studies have shown that the PFC is activated during rule acquisition (that is, rule search and discovery)45. Other imaging studies have indicated activation of the frontopolar cortex during learning of new rules45, as well as during exploration of events and goals in the environment encountered outside the ongoing task5,137,138,139. Monkeys with frontopolar cortex lesions show deficits in detection and learning of new abstract rules but no impairments in implementation of learned rules137. The frontopolar cortex is proposed to have a crucial role in mediating the exploration of alternative rules, goals and contingencies in the environment5,27. Therefore, the PFC, particularly its anterior regions, might be necessary for learning and formation of abstract rules and categories, but not for implementation of previously learned rules (Box 2).

The results of other studies suggest that learning of abstract rules is dependent on the hippocampus and underlying cortex but that, once learned, subsequent performance of rule-based tasks becomes independent of the medial temporal lobe131,132,140. Connections between PFC regions and the striatum are thought to form several complementary and parallel networks facilitating the reciprocal information exchange necessary for learning rules through trial and error. According to this model, slow and goal-directed learning is accomplished in the PFC whereas fast and reinforcement-guided learning occurs in the striatum20.

Rule storage and retrieval

After rules are formed, rule definitions based on instructions, functionality or other features need to be maintained as long-term memories to facilitate their future retrieval for subsequent rule-based action selection and generalization to novel exemplars or situations. Imaging studies in humans have shown that distributed networks are activated when learned rules are used to guide behaviour17,20,76,87,92. Neuronal representations of learned rules have been seen in various PFC regions, the PMC, the parietal cortex and even in sensory association areas21,28,29,31,34,35,38,46,63,67,69,78,99,100,104,105,141. One human imaging study suggested that medial temporal regions might also be involved in rule storage, whereas the VLPFC is typically involved in retrieval of rule information142. These findings indicate that information relating to an abstract rule or category might not necessarily be stored within any one particular brain area, such as the PFC. A cross-species study that examined the role of PFC regions in rule-related behavioural modulation26 showed that humans and monkeys performing a computerized WCST analogue exhibited an inherent behavioural bias to a particular sensory dimension that resulted in improved performance favouring one of the rules (although these biases were in opposite directions in monkeys versus humans). This inherent bias was still present in monkeys after lesioning of the DLPFC, ACC or OFC; however, their rule-shifting ability was impaired. Lesioning of the posterior cingulate cortex, superior DLPFC or frontopolar cortex did not impair either rule-shifting or inherent bias favouring a particular rule in monkeys. These findings indicate that the information necessary to implement a rule and inherent bias towards a particular rule both arise from areas outside the PFC that influence behaviour, whereas dynamic shifts between frequently changing rules require frontal cortical regions (specifically the DLPFC, OFC or ACC)26.

These data suggest that abstract rules might influence various cognitive processes that coherently organize goal-directed behaviours such as perception, memory, attention, emotions, motivation and action selection. Therefore, neural representations of rules might be maintained and retrieved in different neural structures that correspond to and depend on the processing stage and the nature of rule-related modulation95.

Rule-guided behaviour

Abstract rules can guide behaviour in environments that have differing levels of novelty or volatility. Under stable conditions, when no changes occur in task contingencies or environmental features, well-learned and practised rules remain efficient in guiding behaviour and achieving goals. In this scenario, rule-guided behaviour does not depend on the integrity of the PFC. Both humans and monkeys with lesions in the PFC show deficits in shifting between rules in the WCST or WCST analogues but not in implementing individual rules when frequent rule shifts are not required7,28,39,42 (Fig. 5).

By contrast, in new or volatile environments, the validity of learned rules frequently changes and therefore the most suitable rule must be selected for guiding behaviour and achieving goals. Rule selection is achieved by assessing the decision outcome and shifting between rules, which entails abandoning (inhibiting) irrelevant rules and related information to focus, instead, on the relevant rule2,6.

Integration with executive functions

Rule implementation in complex tasks also requires the integration of rules with categories or contextual factors to appropriately guide behaviour. The formation and dynamic updating of abstract rules depend on executive functions that coherently integrate various sources of information10,20. Once learned, abstract rules can influence and direct executive functions to facilitate achievement of goals.

Neuronal activity recordings indicate that, in the context of the WCST, abstract rules are encoded in the PFC within and across trials31,32,33,86. Figure 3a shows the activity of a single DLPFC neuron recorded while a monkey performed a computerized version of the WCST in which abstract rules were frequently changed without any notification, which required the monkey to infer the new rule by trial and error. This neuron showed activity related to the abstract rule even during inter-trial intervals when no task-relevant stimulus was present31. Thus, encoded information reflecting the interaction of abstract rules with executive functions, such as working memory, might influence task preparation106 and guide rule-based decisions in the upcoming trial32,86. Information about the relevant rule might direct selective attention to the currently relevant stimulus dimension, and thereby optimize action selection31,32,33,78,86,108, as well as enabling context-dependent interpretation of the behavioural outcome of the action (that is, use of a correct or incorrect rule) and updating working memory of the rule7,29,78,89,92,108. Figure 3b shows the activity of another single neuron in the DLPFC, recorded in a monkey performing the WCST analogue. A visual error signal was shown when the monkey committed a perseverative error (meaning that a previously relevant but currently irrelevant rule continued to be applied), which mainly occurred immediately after a rule change. Neuronal responses to these error signals were modulated by the abstract rule, indicating that assessment of decision outcome — an important aspect of executive functions — might have been modulated by the abstract rule31. Such interactions between abstract rules and executive functions are crucial for the interpretation of decision outcomes and adaptive behaviour in the WCST; these studies in monkeys indicate that the activity of prefrontal neurons conveys this rule-related information29,31.


The behavioural repertoire of primates is enriched by abstract rules. Accumulating evidence suggests that a distributed constellation of cortical and subcortical brain regions is involved in the formation, maintenance, implementation and updating of these entities6,7,26,95 (Fig. 5). Cross-species studies indicate that abstract rules are formed and used for guiding behaviour in both humans and non-human primates and that homologous brain regions are involved in mediating these processes.

Abstract rules can influence different phases of goal-directed behaviours (Fig. 4). For example, abstract rules might modulate preparations for initiating and performing a given task or task set79,83,88,93,99,106, guide stimulus or feature selection29,31,47,51,63,80,102,143, support conflict resolution3,4,6,28, underlie action selection and response inhibition6,7,25,81,144,145, and influence the assessment of behavioural (decision) outcomes7,29,78. Therefore, formation and goal-directed use of abstract rules are both inherently linked to executive functions. This link might account for findings in both humans and monkeys that show activation of the lateral PFC, ACC and PPC regions when context-dependent implementation of abstract rules is necessary for action selection and achieving goals18,84,85,86,87,88,89,96. These brain regions are the main nodes of a distributed executive control network involved in organizing various goal-directed behaviours89,146. In this view, task context, task demands on executive functions and environmental volatility or stability are the main factors that determine how, when and which cognitive processes are influenced by abstract rules.

Pathway-specific and cell-specific inactivation or modulation of neural activity147,148 might shed additional light on the neural basis of rule formation and implementation in monkeys and bring insight to the neural basis of abstract rules in the human brain. Elucidating the neural substrate and underlying mechanisms of abstract rules and categories is important not only for understanding the neural basis of some of our most advanced cognitive functions but also because this knowledge might bring mechanistic insights to the deficits in rule-based behaviour associated with neuropsychological disorders (Box 1) and their underlying neural malfunctions11,12,13,16.


  1. 1.

    Ashby, F. G. & Ell, S. W. The neurobiology of human category learning. Trends Cogn. Sci. 5, 204–210 (2001).

    PubMed  Google Scholar 

  2. 2.

    Miller, E. K. & Cohen, J. D. An integrative theory of prefrontal cortex function. Annu. Rev. Neurosci. 24, 167–202 (2001).

    CAS  PubMed  Google Scholar 

  3. 3.

    Mansouri, F. A. & Buckley, M. J. Context-dependent adjustments in executive control of goal-directed behaviour: contribution of frontal brain areas to conflict-induced behavioural adjustments in primates. Adv. Neurobiol. 21, 71–83 (2018).

    PubMed  Google Scholar 

  4. 4.

    Mansouri, F. A., Egner, T. & Buckley, M. J. Monitoring demands for executive control: shared functions between human and nonhuman primates. Trends Neurosci. 40, 15–27 (2017).

    CAS  PubMed  Google Scholar 

  5. 5.

    Mansouri, F. A., Koechlin, E., Rosa, M. G. P. & Buckley, M. J. Managing competing goals — a key role for the frontopolar cortex. Nat. Rev. Neurosci. 18, 645–657 (2017). This comprehensive review proposes that the frontopolar cortex is involved in adjusting the balance between exploitation and exploration in primates.

    CAS  PubMed  Google Scholar 

  6. 6.

    Mansouri, F. A., Tanaka, K. & Buckley, M. J. Conflict-induced behavioural adjustment: a clue to the executive functions of the prefrontal cortex. Nat. Rev. Neurosci. 10, 141–152 (2009).

    CAS  PubMed  Google Scholar 

  7. 7.

    Buckley, M. J. et al. Dissociable components of rule-guided behavior depend on distinct medial and prefrontal regions. Science 325, 52–58 (2009). This study examines the effects of selective lesions within different prefrontal regions on cognitive flexibility in shifting between abstract rules.

    CAS  PubMed  Google Scholar 

  8. 8.

    Freedman, D. J. & Assad, J. A. Neuronal mechanisms of visual categorization: an abstract view on decision making. Annu. Rev. Neurosci. 39, 129–147 (2016).

    CAS  PubMed  Google Scholar 

  9. 9.

    Katz, J. S., Wright, A. A. & Bodily, K. D. Issues in the comparative cognition of abstract-concept learning. Comp. Cogn. Behav. Rev. 2, 79–92 (2007).

    PubMed  PubMed Central  Google Scholar 

  10. 10.

    Pan, X. & Sakagami, M. Category representation and generalization in the prefrontal cortex. Eur. J. Neurosci. 35, 1083–1091 (2012).

    PubMed  Google Scholar 

  11. 11.

    Alderson-Day, B. & McGonigle-Chalmers, M. Is it a bird? Is it a plane? Category use in problem-solving in children with autism spectrum disorders. J. Autism Dev. Disord. 41, 555–565 (2011).

    PubMed  Google Scholar 

  12. 12.

    Gastgeb, H. Z., Dundas, E. M., Minshew, N. J. & Strauss, M. S. Category formation in autism: can individuals with autism form categories and prototypes of dot patterns? J. Autism Dev. Disord. 42, 1694–1704 (2012).

    PubMed  PubMed Central  Google Scholar 

  13. 13.

    Jones, E. J., Webb, S. J., Estes, A. & Dawson, G. Rule learning in autism: the role of reward type and social context. Dev. Neuropsychol. 38, 58–77 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  14. 14.

    Keri, S., Kalman, J., Kelemen, O., Benedek, G. & Janka, Z. Are Alzheimer’s disease patients able to learn visual prototypes? Neuropsychologia 39, 1218–1223 (2001).

    CAS  PubMed  Google Scholar 

  15. 15.

    Keri, S. et al. Abstraction is impaired at the perceptual level in schizophrenic patients. Neurosci. Lett. 243, 93–96 (1998).

    CAS  PubMed  Google Scholar 

  16. 16.

    Yerys, B. E. et al. Neural correlates of set-shifting in children with autism. Autism Res. 8, 386–397 (2015).

    PubMed  PubMed Central  Google Scholar 

  17. 17.

    Bunge, S. A. How we use rules to select actions: a review of evidence from cognitive neuroscience. Cogn. Affect. Behav. Neurosci. 4, 564–579 (2004).

    PubMed  Google Scholar 

  18. 18.

    Koechlin, E., Ody, C. & Kouneiher, F. The architecture of cognitive control in the human prefrontal cortex. Science 302, 1181–1185 (2003).

    CAS  PubMed  Google Scholar 

  19. 19.

    Koechlin, E. & Summerfield, C. An information theoretical approach to prefrontal executive function. Trends Cogn. Sci. 11, 229–235 (2007).

    PubMed  Google Scholar 

  20. 20.

    Seger, C. A. & Miller, E. K. Category learning in the brain. Annu. Rev. Neurosci. 33, 203–219 (2010).

    CAS  PubMed  PubMed Central  Google Scholar 

  21. 21.

    Viswanathan, P. & Nieder, A. Neuronal correlates of a visual ‘sense of number’ in primate parietal and prefrontal cortices. Proc. Natl Acad. Sci. USA 110, 11187–11192 (2013).

    CAS  PubMed  Google Scholar 

  22. 22.

    Cole, M. W., Etzel, J. A., Zacks, J. M., Schneider, W. & Braver, T. S. Rapid transfer of abstract rules to novel contexts in human lateral prefrontal cortex. Front. Hum. Neurosci. 5, 142 (2011).

    PubMed  PubMed Central  Google Scholar 

  23. 23.

    Buckley, M. J. & Sigala, N. Is top-down control from prefrontal cortex necessary for visual categorization? Neuron 66, 471–473 (2010).

    CAS  PubMed  Google Scholar 

  24. 24.

    Dias, R., Robbins, T. W. & Roberts, A. C. Dissociation in prefrontal cortex of affective and attentional shifts. Nature 380, 69–72 (1996).

    CAS  PubMed  Google Scholar 

  25. 25.

    Hoshi, E., Shima, K. & Tanji, J. Neuronal activity in the primate prefrontal cortex in the process of motor selection based on two behavioral rules. J. Neurophysiol. 83, 2355–2373 (2000).

    CAS  PubMed  Google Scholar 

  26. 26.

    Mansouri, F. A., Buckley, M. J., Fehring, D. J. & Tanaka, K. The role of primate prefrontal cortex in bias and shift between visual dimensions. Cereb. Cortex 30, 85–99 (2020). This cross-species study reveals that both monkeys and humans show bias to a particular dimension in the WCST and that such biases are not dependent on the PFC.

    PubMed  Google Scholar 

  27. 27.

    Mansouri, F. A., Buckley, M. J., Mahboubi, M. & Tanaka, K. Behavioral consequences of selective damage to frontal pole and posterior cingulate cortices. Proc. Natl Acad. Sci. USA 112, E3940–E3949 (2015).

    CAS  PubMed  Google Scholar 

  28. 28.

    Mansouri, F. A., Buckley, M. J. & Tanaka, K. Mnemonic function of the dorsolateral prefrontal cortex in conflict-induced behavioral adjustment. Science 318, 987–990 (2007).

    CAS  PubMed  Google Scholar 

  29. 29.

    Mansouri, F. A., Buckley, M. J. & Tanaka, K. The essential role of primate orbitofrontal cortex in conflict-induced executive control adjustment. J. Neurosci. 34, 11016–11031 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  30. 30.

    Mansouri, F. A., Fehring, D. J., Gaillard, A., Jaberzadeh, S. & Parkington, H. Sex dependency of inhibitory control functions. Biol. Sex. Differ. 7, 11 (2016).

    PubMed  PubMed Central  Google Scholar 

  31. 31.

    Mansouri, F. A., Matsumoto, K. & Tanaka, K. Prefrontal cell activities related to monkeys’ success and failure in adapting to rule changes in a Wisconsin Card Sorting Test analog. J. Neurosci. 26, 2745–2756 (2006).

    CAS  PubMed  PubMed Central  Google Scholar 

  32. 32.

    Mansouri, F. A., Rosa, M. G. & Atapour, N. Working memory in the service of executive control functions. Front. Syst. Neurosci. 9, 166 (2015).

    PubMed  PubMed Central  Google Scholar 

  33. 33.

    Mansouri, F. A. & Tanaka, K. Behavioral evidence for working memory of sensory dimension in macaque monkeys. Behav. Brain Res. 136, 415–426 (2002).

    PubMed  Google Scholar 

  34. 34.

    Wallis, J. D., Anderson, K. C. & Miller, E. K. Single neurons in prefrontal cortex encode abstract rules. Nature 411, 953–956 (2001). This paper shows how the activity of single neurons in the monkey PFC conveys information about abstract rules.

    CAS  PubMed  Google Scholar 

  35. 35.

    Wallis, J. D. & Miller, E. K. From rule to response: neuronal processes in the premotor and prefrontal cortex. J. Neurophysiol. 90, 1790–1806 (2003).

    PubMed  Google Scholar 

  36. 36.

    Nieder, A. Counting on neurons: the neurobiology of numerical competence. Nat. Rev. Neurosci. 6, 177–190 (2005).

    CAS  PubMed  Google Scholar 

  37. 37.

    Nieder, A. & Dehaene, S. Representation of number in the brain. Annu. Rev. Neurosci. 32, 185–208 (2009).

    CAS  PubMed  Google Scholar 

  38. 38.

    Nieder, A., Freedman, D. J. & Miller, E. K. Representation of the quantity of visual items in the primate prefrontal cortex. Science 297, 1708–1711 (2002). This pioneering study shows how single neurons in the monkey PFC encode abstract numerical information.

    CAS  PubMed  Google Scholar 

  39. 39.

    Drewe, E. A. The effect of type and area of brain lesion on Wisconsin Card Sorting Test performance. Cortex 10, 159–170 (1974).

    CAS  PubMed  Google Scholar 

  40. 40.

    Milner, B. Effects of different brain lesions on card sorting. Arch. Neurol. 9, 90–100 (1963).

    Google Scholar 

  41. 41.

    Stuss, D. T. et al. The involvement of orbitofrontal cerebrum in cognitive tasks. Neuropsychologia 21, 235–248 (1983).

    CAS  PubMed  Google Scholar 

  42. 42.

    Stuss, D. T. et al. Wisconsin Card Sorting Test performance in patients with focal frontal and posterior brain damage: effects of lesion location and test structure on separable cognitive processes. Neuropsychologia 38, 388–402 (2000).

    CAS  PubMed  Google Scholar 

  43. 43.

    La Camera, G., Bouret, S. & Richmond, B. J. Contributions of lateral and orbital frontal regions to abstract rule acquisition and reversal in monkeys. Front. Neurosci. 12, 165 (2018).

    PubMed  PubMed Central  Google Scholar 

  44. 44.

    Matsumoto, N., Eldridge, M. A., Saunders, R. C., Reoli, R. & Richmond, B. J. Mild perceptual categorization deficits follow bilateral removal of anterior inferior temporal cortex in rhesus monkeys. J. Neurosci. 36, 43–53 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  45. 45.

    Crescentini, C. et al. Mechanisms of rule acquisition and rule following in inductive reasoning. J. Neurosci. 31, 7763–7774 (2011).

    CAS  PubMed  PubMed Central  Google Scholar 

  46. 46.

    Miller, E. K., Nieder, A., Freedman, D. J. & Wallis, J. D. Neural correlates of categories and concepts. Curr. Opin. Neurobiol. 13, 198–203 (2003).

    CAS  PubMed  Google Scholar 

  47. 47.

    Bunge, S. A. & Wallis, J. D. Neuroscience of Rule-Guided Behavior (Oxford Univ. Press, 2008).

  48. 48.

    Lazarowski, L., Goodman, A., Galizio, M. & Bruce, K. Effects of set size on identity and oddity abstract-concept learning in rats. Anim. Cogn. 22, 733–742 (2019).

    PubMed  Google Scholar 

  49. 49.

    Smith, J. D. et al. Generalization of category knowledge and dimensional categorization in humans (Homo sapiens) and nonhuman primates (Macaca mulatta). J. Exp. Psychol. Anim. Learn. Cogn. 41, 322–335 (2015).

    PubMed  PubMed Central  Google Scholar 

  50. 50.

    Freedman, D. J. & Assad, J. A. Experience-dependent representation of visual categories in parietal cortex. Nature 443, 85–88 (2006).

    CAS  PubMed  Google Scholar 

  51. 51.

    Freedman, D. J. & Miller, E. K. Neural mechanisms of visual categorization: insights from neurophysiology. Neurosci. Biobehav. Rev. 32, 311–329 (2008).

    PubMed  Google Scholar 

  52. 52.

    Freedman, D. J., Riesenhuber, M., Poggio, T. & Miller, E. K. Categorical representation of visual stimuli in the primate prefrontal cortex. Science 291, 312–316 (2001). This pioneering study shows how single neurons in the monkey PFC encode multifaceted categories.

    CAS  PubMed  Google Scholar 

  53. 53.

    Baugh, A. T., Akre, K. L. & Ryan, M. J. Categorical perception of a natural, multivariate signal: mating call recognition in tungara frogs. Proc. Natl Acad. Sci. USA 105, 8985–8988 (2008).

    CAS  PubMed  Google Scholar 

  54. 54.

    Veit, L. & Nieder, A. Abstract rule neurons in the endbrain support intelligent behaviour in corvid songbirds. Nat. Commun. 4, 2878 (2013).

    PubMed  Google Scholar 

  55. 55.

    Wyttenbach, R. A., May, M. L. & Hoy, R. R. Categorical perception of sound frequency by crickets. Science 273, 1542–1544 (1996). This pioneering study provides solid evidence for category-guided behaviour in insects.

    CAS  PubMed  Google Scholar 

  56. 56.

    Giurfa, M., Zhang, S., Jenett, A., Menzel, R. & Srinivasan, M. V. The concepts of ‘sameness’ and ‘difference’ in an insect. Nature 410, 930–933 (2001).

    CAS  PubMed  Google Scholar 

  57. 57.

    DeGutis, J. & D’Esposito, M. Distinct mechanisms in visual category learning. Cogn. Affect. Behav. Neurosci. 7, 251–259 (2007).

    PubMed  Google Scholar 

  58. 58.

    Freedman, D. J., Riesenhuber, M., Poggio, T. & Miller, E. K. A comparison of primate prefrontal and inferior temporal cortices during visual categorization. J. Neurosci. 23, 5235–5246 (2003).

    CAS  PubMed  PubMed Central  Google Scholar 

  59. 59.

    Mahon, B. Z. & Caramazza, A. Concepts and categories: a cognitive neuropsychological perspective. Annu. Rev. Psychol. 60, 27–51 (2009).

    PubMed  PubMed Central  Google Scholar 

  60. 60.

    Sigala, N. & Logothetis, N. K. Visual categorization shapes feature selectivity in the primate temporal cortex. Nature 415, 318–320 (2002).

    CAS  PubMed  Google Scholar 

  61. 61.

    ten Cate, C. & Okanoya, K. Revisiting the syntactic abilities of non-human animals: natural vocalizations and artificial grammar learning. Philos. Trans. R. Soc. Lond. B Biol. Sci. 367, 1984–1994 (2012).

    PubMed  PubMed Central  Google Scholar 

  62. 62.

    Tanji, J., Shima, K. & Mushiake, H. Concept-based behavioral planning and the lateral prefrontal cortex. Trends Cogn. Sci. 11, 528–534 (2007).

    PubMed  Google Scholar 

  63. 63.

    Muhammad, R., Wallis, J. D. & Miller, E. K. A comparison of abstract rules in the prefrontal cortex, premotor cortex, inferior temporal cortex, and striatum. J. Cogn. Neurosci. 18, 974–989 (2006).

    PubMed  Google Scholar 

  64. 64.

    Bunge, S. A., Kahn, I., Wallis, J. D., Miller, E. K. & Wagner, A. D. Neural circuits subserving the retrieval and maintenance of abstract rules. J. Neurophysiol. 90, 3419–3428 (2003).

    PubMed  Google Scholar 

  65. 65.

    Elliott, R. & Dolan, R. J. Differential neural responses during performance of matching and nonmatching to sample tasks at two delay intervals. J. Neurosci. 19, 5066–5073 (1999).

    CAS  PubMed  PubMed Central  Google Scholar 

  66. 66.

    Balsters, J. H., Whelan, C. D., Robertson, I. H. & Ramnani, N. Cerebellum and cognition: evidence for the encoding of higher order rules. Cereb. Cortex 23, 1433–1443 (2013).

    PubMed  Google Scholar 

  67. 67.

    Bongard, S. & Nieder, A. Basic mathematical rules are encoded by primate prefrontal cortex neurons. Proc. Natl Acad. Sci. USA 107, 2277–2282 (2010).

    CAS  PubMed  Google Scholar 

  68. 68.

    Eiselt, A. K. & Nieder, A. Representation of abstract quantitative rules applied to spatial and numerical magnitudes in primate prefrontal cortex. J. Neurosci. 33, 7526–7534 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  69. 69.

    Eiselt, A. K. & Nieder, A. Rule activity related to spatial and numerical magnitudes: comparison of prefrontal, premotor, and cingulate motor cortices. J. Cogn. Neurosci. 26, 1000–1012 (2014).

    PubMed  Google Scholar 

  70. 70.

    Kutter, E. F., Bostroem, J., Elger, C. E., Mormann, F. & Nieder, A. Single neurons in the human brain encode numbers. Neuron 100, 753–761.e4 (2018).

    CAS  PubMed  Google Scholar 

  71. 71.

    Nieder, A. Supramodal numerosity selectivity of neurons in primate prefrontal and posterior parietal cortices. Proc. Natl Acad. Sci. USA 109, 11860–11865 (2012).

    CAS  PubMed  Google Scholar 

  72. 72.

    Vallentin, D., Bongard, S. & Nieder, A. Numerical rule coding in the prefrontal, premotor, and posterior parietal cortices of macaques. J. Neurosci. 32, 6621–6630 (2012).

    CAS  PubMed  PubMed Central  Google Scholar 

  73. 73.

    Daitch, A. L. et al. Mapping human temporal and parietal neuronal population activity and functional coupling during mathematical cognition. Proc. Natl Acad. Sci. USA 113, E7277–E7286 (2016).

    CAS  PubMed  Google Scholar 

  74. 74.

    Cromer, J. A., Roy, J. E. & Miller, E. K. Representation of multiple, independent categories in the primate prefrontal cortex. Neuron 66, 796–807 (2010). This pioneering study shows a dynamic training-dependent representation of category information in PFC neuronal activity.

    CAS  PubMed  PubMed Central  Google Scholar 

  75. 75.

    Forstmann, B. U., Brass, M., Koch, I. & von Cramon, D. Y. Internally generated and directly cued task sets: an investigation with fMRI. Neuropsychologia 43, 943–952 (2005).

    PubMed  Google Scholar 

  76. 76.

    Bengtsson, S. L., Haynes, J. D., Sakai, K., Buckley, M. J. & Passingham, R. E. The representation of abstract task rules in the human prefrontal cortex. Cereb. Cortex 19, 1929–1936 (2009).

    PubMed  Google Scholar 

  77. 77.

    Mian, M. K. et al. Encoding of rules by neurons in the human dorsolateral prefrontal cortex. Cereb. Cortex 24, 807–816 (2014).

    PubMed  Google Scholar 

  78. 78.

    Kuwabara, M., Mansouri, F. A., Buckley, M. J. & Tanaka, K. Cognitive control functions of anterior cingulate cortex in macaque monkeys performing a Wisconsin Card Sorting Test analog. J. Neurosci. 34, 7531–7547 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  79. 79.

    Kamigaki, T., Fukushima, T. & Miyashita, Y. Cognitive set reconfiguration signaled by macaque posterior parietal neurons. Neuron 61, 941–951 (2009).

    CAS  PubMed  Google Scholar 

  80. 80.

    Kamigaki, T., Fukushima, T., Tamura, K. & Miyashita, Y. Neurodynamics of cognitive set shifting in monkey frontal cortex and its causal impact on behavioral flexibility. J. Cogn. Neurosci. 24, 2171–2185 (2012).

    PubMed  Google Scholar 

  81. 81.

    Sleezer, B. J., Castagno, M. D. & Hayden, B. Y. Rule encoding in orbitofrontal cortex and striatum guides selection. J. Neurosci. 36, 11223–11237 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  82. 82.

    Sleezer, B. J. & Hayden, B. Y. Differential contributions of ventral and dorsal striatum to early and late phases of cognitive set reconfiguration. J. Cogn. Neurosci. 28, 1849–1864 (2016).

    PubMed  PubMed Central  Google Scholar 

  83. 83.

    Sleezer, B. J., LoConte, G. A., Castagno, M. D. & Hayden, B. Y. Neuronal responses support a role for orbitofrontal cortex in cognitive set reconfiguration. Eur. J. Neurosci. 45, 940–951 (2017).

    PubMed  PubMed Central  Google Scholar 

  84. 84.

    Nakahara, K., Hayashi, T., Konishi, S. & Miyashita, Y. Functional MRI of macaque monkeys performing a cognitive set-shifting task. Science 295, 1532–1536 (2002). This pioneering fMRI study in monkeys and humans reveals how shifting between rules is represented in the PFC.

    CAS  PubMed  Google Scholar 

  85. 85.

    Petrides, M., Tomaiuolo, F., Yeterian, E. H. & Pandya, D. N. The prefrontal cortex: comparative architectonic organization in the human and the macaque monkey brains. Cortex 48, 46–57 (2012).

    PubMed  Google Scholar 

  86. 86.

    Konishi, S. et al. Contribution of working memory to transient activation in human inferior prefrontal cortex during performance of the Wisconsin Card Sorting Test. Cereb. Cortex 9, 745–753 (1999).

    CAS  PubMed  Google Scholar 

  87. 87.

    Konishi, S. et al. Transient activation of inferior prefrontal cortex during cognitive set shifting. Nat. Neurosci. 1, 80–84 (1998).

    CAS  PubMed  Google Scholar 

  88. 88.

    Asari, T., Konishi, S., Jimura, K. & Miyashita, Y. Multiple components of lateral posterior parietal activation associated with cognitive set shifting. Neuroimage 26, 694–702 (2005).

    PubMed  Google Scholar 

  89. 89.

    Buchsbaum, B. R., Greer, S., Chang, W. L. & Berman, K. F. Meta-analysis of neuroimaging studies of the Wisconsin card-sorting task and component processes. Hum. Brain Mapp. 25, 35–45 (2005).

    PubMed  PubMed Central  Google Scholar 

  90. 90.

    Menon, V. & Uddin, L. Q. Saliency, switching, attention and control: a network model of insula function. Brain Struct. Funct. 214, 655–667 (2010).

    PubMed  PubMed Central  Google Scholar 

  91. 91.

    Mentzel, H. J. et al. Cognitive stimulation with the Wisconsin Card Sorting Test: functional MR imaging at 1.5T. Radiology 207, 399–404 (1998).

    CAS  PubMed  Google Scholar 

  92. 92.

    Monchi, O., Petrides, M., Petre, V., Worsley, K. & Dagher, A. Wisconsin card sorting revisited: distinct neural circuits participating in different stages of the task identified by event-related functional magnetic resonance imaging. J. Neurosci. 21, 7733–7741 (2001).

    CAS  PubMed  PubMed Central  Google Scholar 

  93. 93.

    Ravizza, S. M. & Carter, C. S. Shifting set about task switching: behavioral and neural evidence for distinct forms of cognitive flexibility. Neuropsychologia 46, 2924–2935 (2008).

    PubMed  PubMed Central  Google Scholar 

  94. 94.

    Wang, J., Cao, B., Cai, X., Gao, H. & Li, F. Brain activation of negative feedback in rule acquisition revealed in a segmented Wisconsin Card Sorting Test. PLoS ONE 10, e0140731 (2015).

    PubMed  PubMed Central  Google Scholar 

  95. 95.

    Zeithamova, D. et al. Brain mechanisms of concept learning. J. Neurosci. 39, 8259–8266 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  96. 96.

    Konishi, S. et al. Hemispheric asymmetry in human lateral prefrontal cortex during cognitive set shifting. Proc. Natl Acad. Sci. USA 99, 7803–7808 (2002).

    CAS  PubMed  Google Scholar 

  97. 97.

    Genovesio, A., Brasted, P. J., Mitz, A. R. & Wise, S. P. Prefrontal cortex activity related to abstract response strategies. Neuron 47, 307–320 (2005).

    CAS  PubMed  PubMed Central  Google Scholar 

  98. 98.

    Bussey, T. J., Wise, S. P. & Murray, E. A. The role of ventral and orbital prefrontal cortex in conditional visuomotor learning and strategy use in rhesus monkeys (Macaca mulatta). Behav. Neurosci. 115, 971–982 (2001).

    CAS  PubMed  Google Scholar 

  99. 99.

    Stoet, G. & Snyder, L. H. Single neurons in posterior parietal cortex of monkeys encode cognitive set. Neuron 42, 1003–1012 (2004).

    CAS  PubMed  Google Scholar 

  100. 100.

    Stoet, G. & Snyder, L. H. Neural correlates of executive control functions in the monkey. Trends Cogn. Sci. 13, 228–234 (2009).

    PubMed  PubMed Central  Google Scholar 

  101. 101.

    White, I. M. & Wise, S. P. Rule-dependent neuronal activity in the prefrontal cortex. Exp. Brain Res. 126, 315–335 (1999).

    CAS  PubMed  Google Scholar 

  102. 102.

    Asaad, W. F., Rainer, G. & Miller, E. K. Task-specific neural activity in the primate prefrontal cortex. J. Neurophysiol. 84, 451–459 (2000).

    CAS  PubMed  Google Scholar 

  103. 103.

    Mante, V., Sussillo, D., Shenoy, K. V. & Newsome, W. T. Context-dependent computation by recurrent dynamics in prefrontal cortex. Nature 503, 78–84 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  104. 104.

    Crowe, D. A. et al. Prefrontal neurons transmit signals to parietal neurons that reflect executive control of cognition. Nat. Neurosci. 16, 1484–1491 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  105. 105.

    Goodwin, S. J., Blackman, R. K., Sakellaridi, S. & Chafee, M. V. Executive control over cognition: stronger and earlier rule-based modulation of spatial category signals in prefrontal cortex relative to parietal cortex. J. Neurosci. 32, 3499–3515 (2012).

    CAS  PubMed  PubMed Central  Google Scholar 

  106. 106.

    Sakai, K. & Passingham, R. E. Prefrontal interactions reflect future task operations. Nat. Neurosci. 6, 75–81 (2003).

    CAS  PubMed  Google Scholar 

  107. 107.

    Swaminathan, S. K. & Freedman, D. J. Preferential encoding of visual categories in parietal cortex compared with prefrontal cortex. Nat. Neurosci. 15, 315–320 (2012).

    CAS  PubMed  PubMed Central  Google Scholar 

  108. 108.

    Lie, C. H., Specht, K., Marshall, J. C. & Fink, G. R. Using fMRI to decompose the neural processes underlying the Wisconsin Card Sorting Test. Neuroimage 30, 1038–1049 (2006).

    PubMed  Google Scholar 

  109. 109.

    Duncan, J. & Owen, A. M. Common regions of the human frontal lobe recruited by diverse cognitive demands. Trends Neurosci. 23, 475–483 (2000). This influential paper proposes that a distributed network including prefrontal, cingulate and parietal cortices supports the organization and execution of various goal-directed behaviours.

    CAS  PubMed  Google Scholar 

  110. 110.

    Brown, V. J. & Tait, D. S. Attentional set-shifting across species. Curr. Top. Behav. Neurosci. 28, 363–395 (2016).

    CAS  PubMed  Google Scholar 

  111. 111.

    Goldman, P. S. & Rosvold, H. E. Localization of function within the dorsolateral prefrontal cortex of the rhesus monkey. Exp. Neurol. 27, 291–304 (1970).

    CAS  PubMed  Google Scholar 

  112. 112.

    Petrides, M. Motor conditional associative-learning after selective prefrontal lesions in the monkey. Behav. Brain Res. 5, 407–413 (1982).

    CAS  PubMed  Google Scholar 

  113. 113.

    Petrides, M. Deficits in non-spatial conditional associative learning after periarcuate lesions in the monkey. Behav. Brain Res. 16, 95–101 (1985).

    CAS  PubMed  Google Scholar 

  114. 114.

    Podbros, L. Z., Stamm, J. S. & Denaro, F. J. Associative function of the arcuate segment of the monkey’s prefrontal cortex. Physiol. Behav. 24, 103–109 (1980).

    CAS  PubMed  Google Scholar 

  115. 115.

    Petrides, M. Deficits on conditional associative-learning tasks after frontal- and temporal-lobe lesions in man. Neuropsychologia 23, 601–614 (1985).

    CAS  PubMed  Google Scholar 

  116. 116.

    Petrides, M. Visuo-motor conditional associative learning after frontal and temporal lesions in the human brain. Neuropsychologia 35, 989–997 (1997).

    CAS  PubMed  Google Scholar 

  117. 117.

    Kowalska, D. M., Bachevalier, J. & Mishkin, M. The role of the inferior prefrontal convexity in performance of delayed nonmatching-to-sample. Neuropsychologia 29, 583–600 (1991).

    CAS  PubMed  Google Scholar 

  118. 118.

    Bachevalier, J. & Mishkin, M. Visual recognition impairment follows ventromedial but not dorsolateral prefrontal lesions in monkeys. Behav. Brain Res. 20, 249–261 (1986).

    CAS  PubMed  Google Scholar 

  119. 119.

    Meunier, M., Bachevalier, J. & Mishkin, M. Effects of orbital frontal and anterior cingulate lesions on object and spatial memory in rhesus monkeys. Neuropsychologia 35, 999–1015 (1997).

    CAS  PubMed  Google Scholar 

  120. 120.

    Bauer, R. H. & Fuster, J. M. Delayed-matching and delayed-response deficit from cooling dorsolateral prefrontal cortex in monkeys. J. Comp. Physiol. Psychol. 90, 293–302 (1976).

    CAS  PubMed  Google Scholar 

  121. 121.

    Elliott, R., Dolan, R. J. & Frith, C. D. Dissociable functions in the medial and lateral orbitofrontal cortex: evidence from human neuroimaging studies. Cereb. Cortex 10, 308–317 (2000).

    CAS  PubMed  Google Scholar 

  122. 122.

    Moore, T. L., Schettler, S. P., Killiany, R. J., Rosene, D. L. & Moss, M. B. Impairment in delayed nonmatching to sample following lesions of dorsal prefrontal cortex. Behav. Neurosci. 126, 772–780 (2012).

    PubMed  PubMed Central  Google Scholar 

  123. 123.

    Rao, S. C., Rainer, G. & Miller, E. K. Integration of what and where in the primate prefrontal cortex. Science 276, 821–824 (1997).

    CAS  PubMed  Google Scholar 

  124. 124.

    Shallice, T. & Burgess, P. W. Deficits in strategy application following frontal lobe damage in man. Brain 114, 727–741 (1991).

    PubMed  Google Scholar 

  125. 125.

    Boschin, E. A., Brkic, M. M., Simons, J. S. & Buckley, M. J. Distinct roles for the anterior cingulate and dorsolateral prefrontal cortices during conflict between abstract rules. Cereb. Cortex 27, 34–45 (2017).

    PubMed  Google Scholar 

  126. 126.

    Glascher, J. et al. Lesion mapping of cognitive control and value-based decision making in the prefrontal cortex. Proc. Natl Acad. Sci. USA 109, 14681–14686 (2012). This paper presents a comprehensive neuropsychological examination of the effects of brain lesions on cognitive flexibility in shifting between abstract rules.

    CAS  PubMed  Google Scholar 

  127. 127.

    Minamimoto, T., Saunders, R. C. & Richmond, B. J. Monkeys quickly learn and generalize visual categories without lateral prefrontal cortex. Neuron 66, 501–507 (2010). This paper reports that large lesions in the PFC do not impair using learned categories or learning new categories.

    CAS  PubMed  Google Scholar 

  128. 128.

    Tanaka, K. Inferotemporal cortex and object vision. Annu. Rev. Neurosci. 19, 109–139 (1996). This comprehensive review describes the architecture of the ventral visual pathway of object recognition in primates.

    CAS  PubMed  Google Scholar 

  129. 129.

    Buckley, M. J. & Gaffan, D. Perirhinal cortical contributions to object perception. Trends Cogn. Sci. 10, 100–107 (2006).

    PubMed  Google Scholar 

  130. 130.

    Cools, R., Clark, L. & Robbins, T. W. Differential responses in human striatum and prefrontal cortex to changes in object and rule relevance. J. Neurosci. 24, 1129–1135 (2004).

    CAS  PubMed  PubMed Central  Google Scholar 

  131. 131.

    Bowman, C. R. & Zeithamova, D. Abstract memory representations in the ventromedial prefrontal cortex and hippocampus support concept generalization. J. Neurosci. 38, 2605–2614 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  132. 132.

    Murray, E. A. & Wise, S. P. Role of the hippocampus plus subjacent cortex but not amygdala in visuomotor conditional learning in rhesus monkeys. Behav. Neurosci. 110, 1261–1270 (1996).

    CAS  PubMed  Google Scholar 

  133. 133.

    Orbach, J., Milner, B. & Rasmussen, T. Learning and retention in monkeys after amygdala–hippocampus resection. Arch. Neurol. 3, 230–251 (1960).

    CAS  PubMed  Google Scholar 

  134. 134.

    Owen, A. M., Roberts, A. C., Polkey, C. E., Sahakian, B. J. & Robbins, T. W. Extra-dimensional versus intra-dimensional set shifting performance following frontal lobe excisions, temporal lobe excisions or amygdalo-hippocampectomy in man. Neuropsychologia 29, 993–1006 (1991).

    CAS  PubMed  Google Scholar 

  135. 135.

    Filoteo, J. V. et al. Cortical and subcortical brain regions involved in rule-based category learning. Neuroreport 16, 111–115 (2005).

    PubMed  Google Scholar 

  136. 136.

    Sloutsky, V. M. From perceptual categories to concepts: what develops? Cogn. Sci. 34, 1244–1286 (2010).

    PubMed  PubMed Central  Google Scholar 

  137. 137.

    Boschin, E. A., Piekema, C. & Buckley, M. J. Essential functions of primate frontopolar cortex in cognition. Proc. Natl Acad. Sci. USA 112, E1020–E1027 (2015).

    CAS  PubMed  Google Scholar 

  138. 138.

    Boorman, E. D., Behrens, T. E. & Rushworth, M. F. Counterfactual choice and learning in a neural network centered on human lateral frontopolar cortex. PLoS Biol. 9, e1001093 (2011).

    CAS  PubMed  PubMed Central  Google Scholar 

  139. 139.

    Daw, N. D., O’Doherty, J. P., Dayan, P., Seymour, B. & Dolan, R. J. Cortical substrates for exploratory decisions in humans. Nature 441, 876–879 (2006).

    CAS  PubMed  PubMed Central  Google Scholar 

  140. 140.

    Wise, S. P. & Murray, E. A. Arbitrary associations between antecedents and actions. Trends Neurosci. 23, 271–276 (2000).

    CAS  PubMed  Google Scholar 

  141. 141.

    Zhou, Y. & Freedman, D. J. Posterior parietal cortex plays a causal role in perceptual and categorical decisions. Science 365, 180–185 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  142. 142.

    Donohue, S. E., Wendelken, C., Crone, E. A. & Bunge, S. A. Retrieving rules for behavior from long-term memory. Neuroimage 26, 1140–1149 (2005).

    PubMed  Google Scholar 

  143. 143.

    Ibos, G. & Freedman, D. J. Dynamic integration of task-relevant visual features in posterior parietal cortex. Neuron 83, 1468–1480 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  144. 144.

    Tremblay, L., Gettner, S. N. & Olson, C. R. Neurons with object-centered spatial selectivity in macaque SEF: do they represent locations or rules? J. Neurophysiol. 87, 333–350 (2002).

    PubMed  Google Scholar 

  145. 145.

    Konishi, S. et al. Common inhibitory mechanism in human inferior prefrontal cortex revealed by event-related functional MRI. Brain 122, 981–991 (1999).

    PubMed  Google Scholar 

  146. 146.

    Niendam, T. A. et al. Meta-analytic evidence for a superordinate cognitive control network subserving diverse executive functions. Cogn. Affect. Behav. Neurosci. 12, 241–268 (2012).

    PubMed  PubMed Central  Google Scholar 

  147. 147.

    Eldridge, M. A. et al. Chemogenetic disconnection of monkey orbitofrontal and rhinal cortex reversibly disrupts reward value. Nat. Neurosci. 19, 37–39 (2016).

    CAS  PubMed  Google Scholar 

  148. 148.

    Whissell, P. D., Tohyama, S. & Martin, L. J. The use of DREADDs to deconstruct behavior. Front. Genet. 7, 70 (2016).

    PubMed  PubMed Central  Google Scholar 

  149. 149.

    Reed, P., Watts, H. & Truzoli, R. Flexibility in young people with autism spectrum disorders on a card sort task. Autism 17, 162–171 (2013).

    PubMed  Google Scholar 

  150. 150.

    Luck, S. J. & Gold, J. M. The construct of attention in schizophrenia. Biol. Psychiatry 64, 34–39 (2008).

    PubMed  PubMed Central  Google Scholar 

  151. 151.

    Barbalat, G., Chambon, V., Franck, N., Koechlin, E. & Farrer, C. Organization of cognitive control within the lateral prefrontal cortex in schizophrenia. Arch. Gen. Psychiatry 66, 377–386 (2009).

    PubMed  Google Scholar 

  152. 152.

    den Braber, A. et al. Brain activation during cognitive planning in twins discordant or concordant for obsessive–compulsive symptoms. Brain 133, 3123–3140 (2010).

    Google Scholar 

  153. 153.

    Badre, D. & D’Esposito, M. Is the rostro-caudal axis of the frontal lobe hierarchical? Nat. Rev. Neurosci. 10, 659–669 (2009).

    CAS  PubMed  PubMed Central  Google Scholar 

  154. 154.

    Badre, D. & Nee, D. E. Frontal cortex and the hierarchical control of behavior. Trends Cogn. Sci. 22, 170–188 (2018).

    PubMed  Google Scholar 

  155. 155.

    Bahlmann, J., Blumenfeld, R. S. & D’Esposito, M. The rostro-caudal axis of frontal cortex is sensitive to the domain of stimulus information. Cereb. Cortex 25, 1815–1826 (2015).

    PubMed  Google Scholar 

  156. 156.

    Wendelken, C., Chung, D. & Bunge, S. A. Rostrolateral prefrontal cortex: domain-general or domain-sensitive? Hum. Brain Mapp. 33, 1952–1963 (2012).

    PubMed  Google Scholar 

  157. 157.

    Pischedda, D., Gorgen, K., Haynes, J. D. & Reverberi, C. Neural representations of hierarchical rule sets: the human control system represents rules irrespective of the hierarchical level to which they belong. J. Neurosci. 37, 12281–12296 (2017).

    CAS  PubMed  PubMed Central  Google Scholar 

  158. 158.

    Tsujimoto, S., Genovesio, A. & Wise, S. P. Evaluating self-generated decisions in frontal pole cortex of monkeys. Nat. Neurosci. 13, 120–126 (2010). This pioneering study describes the activity of frontopolar cortex cells in monkeys performing cognitive tasks.

    CAS  PubMed  Google Scholar 

Download references


The authors thank K. Tanaka (RIKEN Centre for Brain Science, Japan) for his contribution to our proposed model of the interaction between abstract rules and executive functions. The authors also thank the Australian Research Council (ARC) Centre of Excellence in Integrative Brain Function; the authors’ research work is partially funded by an ARC Discovery project grant to F.A.M. and a UK Medical Research Project grant to M.J.B.

Author information




The authors contributed equally to all aspects of the article.

Corresponding author

Correspondence to Farshad Alizadeh Mansouri.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information

Nature Reviews Neuroscience thanks S. Bunge, G. Ashby and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.



The cognitive ability to consider other goals and resources, such as reward, outside the ongoing task.

Executive functions

Brain mechanisms that organize and optimize the use of cognitive resources to achieve a goal.


The cognitive ability to optimize gain and decrease cost while performing an ongoing task.

Selective attention

Neural mechanisms involved in focusing cognitive resources on task-relevant sensory-perceptual processes and inhibiting goal-irrelevant stimuli to facilitate achieving goals.

Extra-dimensional shift

In the context of object-discrimination tasks, individuals learn to select an object on the basis of a feature in a specific dimension that is reinforced by a reward; a shift in reinforcement to another feature in a different dimension (for example, from red colour to triangle shape) means that individuals need to shift their choice accordingly to get the reward.

Intra-dimensional shift

In the context of object-discrimination tasks, individuals learn to select an object on the basis of a feature in a specific dimension that is reinforced by a reward; a shift in reinforcement to another feature in the same dimension (for example, from red to blue colour) means that individuals need to shift their choice accordingly to get the reward.

Stimulus–reward reversal

In the context of object-discrimination tasks, the object–reward association contingency of two objects is reversed; individuals must learn to select the currently rewarded object, which was previously the unrewarded object.

Conflict resolution

Achieving goals in cognitive tasks might require resolution of a conflict (competition) between two sources of information or between two opposing responses.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Mansouri, F.A., Freedman, D.J. & Buckley, M.J. Emergence of abstract rules in the primate brain. Nat Rev Neurosci 21, 595–610 (2020).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing