Face-selective cells in the temporal cortex of monkeys

Learning what matters: Synaptic plasticity with invariance to second-order input correlations

Cortical populations of neurons develop sparse representations adapted to the statistics of the environment. To learn efficient population codes, synaptic plasticity mechanisms must differentiate relevant latent features from spurious input correlations, which are omnipresent in cortical networks. Here, we develop a theory for sparse coding and synaptic plasticity that is invariant to second-order correlations in the input. Going beyond classical Hebbian learning, our learning objective explains the functional form of observed excitatory plasticity mechanisms, showing how Hebbian long-term depression (LTD) cancels the sensitivity to second-order correlations so that receptive fields become aligned with features hidden in higher-order statistics. Invariance to second-order correlations enhances the versatility of biologically realistic learning models, supporting optimal decoding from noisy inputs and sparse population coding from spatially correlated stimuli. In a spiking model with triplet spike-timing-dependent plasticity (STDP), we show that individual neurons can learn localized oriented receptive fields, circumventing the need for input preprocessing, such as whitening, or population-level lateral inhibition. The theory advances our understanding of local unsupervised learning in cortical circuits, offers new interpretations of the Bienenstock-Cooper-Munro and triplet STDP models, and assigns a specific functional role to synaptic LTD mechanisms in pyramidal neurons.

The influence of stereopsis on visual saliency in a proto-object based model of selective attention

Some animals including humans use stereoscopic vision which reconstructs spatial information about the environment from the disparity between images captured by eyes in two separate adjacent locations. Like other sensory information, such stereoscopic information is expected to influence attentional selection. We develop a biologically plausible model of binocular vision to study its effect on bottom-up visual attention, i.e., visual saliency. In our model, the scene is organized in terms of proto-objects on which attention acts, rather than on unbound sets of elementary features. We show that taking into account the stereoscopic information improves the performance of the model in the prediction of human eye movements with statistically significant differences.

Modeling Biological Face Recognition with Deep Convolutional Neural Networks

Deep convolutional neural networks (DCNNs) have become the state-of-the-art computational models of biological object recognition. Their remarkable success has helped vision science break new ground, and recent efforts have started to transfer this achievement to research on biological face recognition. In this regard, face detection can be investigated by comparing face-selective biological neurons and brain areas to artificial neurons and model layers. Similarly, face identification can be examined by comparing in vivo and in silico multidimensional "face spaces." In this review, we summarize the first studies that use DCNNs to model biological face recognition. On the basis of a broad spectrum of behavioral and computational evidence, we conclude that DCNNs are useful models that closely resemble the general hierarchical organization of face recognition in the ventral visual pathway and the core face network. In two exemplary spotlights, we emphasize the unique scientific contributions of these models. First, studies on face detection in DCNNs indicate that elementary face selectivity emerges automatically through feedforward processing even in the absence of visual experience. Second, studies on face identification in DCNNs suggest that identity-specific experience and generative mechanisms facilitate this particular challenge. Taken together, as this novel modeling approach enables close control of predisposition (i.e., architecture) and experience (i.e., training data), it may be suited to inform long-standing debates on the substrates of biological face recognition.

The Role of Visual Factors in Dyslexia

What are the causes of dyslexia? Decades of research reflect a determined search for a single cause where a common assumption is that dyslexia is a consequence of problems with converting phonological information into lexical codes. But reading is a highly complex activity requiring many well-functioning mechanisms, and several different visual problems have been documented in dyslexic readers. We critically review evidence from various sources for the role of visual factors in dyslexia, from magnocellular dysfunction through accounts based on abnormal eye movements and attentional processing, to recent proposals that problems with high-level vision contribute to dyslexia. We believe that the role of visual problems in dyslexia has been underestimated in the literature, to the detriment of the understanding and treatment of the disorder. We propose that rather than focusing on a single core cause, the role of visual factors in dyslexia fits well with risk and resilience models that assume that several variables interact throughout prenatal and postnatal development to either promote or hinder efficient reading.

Towards an optimization of functional localizers in non-human primate neuroimaging with (fMRI) frequency-tagging

Non-human primate (NHP) neuroimaging can provide essential insights into the neural basis of human cognitive functions. While functional magnetic resonance imaging (fMRI) localizers can play an essential role in reaching this objective (Russ et al., 2021), they often differ substantially across species in terms of paradigms, measured signals, and data analysis, biasing the comparisons. Here we introduce a functional frequency-tagging face localizer for NHP imaging, successfully developed in humans and outperforming standard face localizers (Gao et al., 2018). FMRI recordings were performed in two awake macaques. Within a rapid 6 Hz stream of natural non-face objects images, human or monkey face stimuli were presented in bursts every 9 s. We also included control conditions with phase-scrambled versions of all images. As in humans, face-selective activity was objectively identified and quantified at the peak of the face-stimulation frequency (0.111 Hz) and its second harmonic (0.222 Hz) in the Fourier domain. Focal activations with a high signal-to-noise ratio were observed in regions previously described as face-selective, mainly in the STS (clusters PL, ML, MF; also, AL, AF), both for human and monkey faces. Robust face-selective activations were also found in the prefrontal cortex of one monkey (PVL and PO clusters). Face-selective neural activity was highly reliable and excluded all contributions from low-level visual cues contained in the amplitude spectrum of the stimuli. These observations indicate that fMRI frequency-tagging provides a highly valuable approach to objectively compare human and monkey visual recognition systems within the same framework.

Intracerebral Electrophysiological Recordings to Understand the Neural Basis of Human Face Recognition

Understanding how the human brain recognizes faces is a primary scientific goal in cognitive neuroscience. Given the limitations of the monkey model of human face recognition, a key approach in this endeavor is the recording of electrophysiological activity with electrodes implanted inside the brain of human epileptic patients. However, this approach faces a number of challenges that must be overcome for meaningful scientific knowledge to emerge. Here we synthesize a 10 year research program combining the recording of intracerebral activity (StereoElectroEncephaloGraphy, SEEG) in the ventral occipito-temporal cortex (VOTC) of large samples of participants and fast periodic visual stimulation (FPVS), to objectively define, quantify, and characterize the neural basis of human face recognition. These large-scale studies reconcile the wide distribution of neural face recognition activity with its (right) hemispheric and regional specialization and extend face-selectivity to anterior regions of the VOTC, including the ventral anterior temporal lobe (VATL) typically affected by magnetic susceptibility artifacts in functional magnetic resonance imaging (fMRI). Clear spatial dissociations in category-selectivity between faces and other meaningful stimuli such as landmarks (houses, medial VOTC regions) or written words (left lateralized VOTC) are found, confirming and extending neuroimaging observations while supporting the validity of the clinical population tested to inform about normal brain function. The recognition of face identity - arguably the ultimate form of recognition for the human brain - beyond mere differences in physical features is essentially supported by selective populations of neurons in the right inferior occipital gyrus and the lateral portion of the middle and anterior fusiform gyrus. In addition, low-frequency and high-frequency broadband iEEG signals of face recognition appear to be largely concordant in the human association cortex. We conclude by outlining the challenges of this research program to understand the neural basis of human face recognition in the next 10 years.

Temporopolar regions of the human brain

Following prolonged neglect during the formative decades of behavioural neurology, the temporopolar region has become a site of vibrant research on the neurobiology of cognition and conduct. This turnaround can be attributed to increasing recognition of neurodegenerative diseases that target temporopolar regions for peak destruction. The resultant syndromes include behavioural dementia, associative agnosia, semantic forms of primary progressive aphasia and semantic dementia. Clinicopathological correlations show that object naming and word comprehension are critically dependent on the language-dominant (usually left) temporopolar region, whereas behavioural control and non-verbal object recognition display a more bilateral representation with a rightward bias. Neuroanatomical experiments in macaques and neuroimaging in humans show that the temporoparietal region sits at the confluence of auditory, visual and limbic streams of processing at the downstream (deep) pole of the 'what' pathway. The functional neuroanatomy of this region revolves around three axes, an anterograde horizontal axis from unimodal to heteromodal and paralimbic cortex; a radial axis where visual (ventral), auditory (dorsal) and paralimbic (medial) territories encircle temporopolar cortex and display hemispheric asymmetry; and a vertical depth-of-processing axis for the associative elaboration of words, objects and interoceptive states. One function of this neural matrix is to support the transformation of object and word representations from unimodal percepts to multimodal concepts. The underlying process is likely to start at canonical gateways that successively lead to generic (superordinate), specific (basic) and unique levels of recognition. A first sign of left temporopolar dysfunction takes the form of taxonomic blurring where boundaries among categories are preserved but not boundaries among exemplars of a category. Semantic paraphasias and coordinate errors in word-picture verification tests are consequences of this phenomenon. Eventually, boundaries among categories are also blurred and comprehension impairments become more profound. The medial temporopolar region belongs to the amygdalocentric component of the limbic system and stands to integrate exteroceptive information with interoceptive states underlying social interactions. Review of the pertinent literature shows that word comprehension and conduct impairments caused by temporopolar strokes and temporal lobectomy are far less severe than those seen in temporopolar atrophies. One explanation for this unexpected discrepancy invokes the miswiring of residual temporopolar neurons during the many years of indolently progressive neurodegeneration. According to this hypothesis, the temporopolar regions become not only dysfunctional but also sources of aberrant outputs that interfere with the function of areas elsewhere in the language and paralimbic networks, a juxtaposition not seen in lobectomy or stroke.

Contributions of low- and high-level contextual mechanisms to human face perception

Contextual modulations at primary stages of visual processing depend on the strength of local input. Contextual modulations at high-level stages of (face) processing show a similar dependence to local input strength. Namely, the discriminability of a facial feature determines the amount of influence of the face context on that feature. How high-level contextual modulations emerge from primary mechanisms is unclear due to the scarcity of empirical research systematically addressing the functional link between the two. We tested (62) young adults' ability to process local input independent of the context using contrast detection and (upright and inverted) morphed facial feature matching tasks. We first investigated contextual modulation magnitudes across tasks to address their shared variance. A second analysis focused on the profile of performance across contextual conditions. In upright eye matching and contrast detection tasks, contextual modulations only correlated at the level of their profile (averaged Fisher-Z transformed r = 1.18, BF10 > 100), but not magnitude (r = .15, BF10 = .61), suggesting the functional independence but similar working principles of the mechanisms involved. Both the profile (averaged Fisher-Z transformed r = .32, BF10 = 9.7) and magnitude (r = .28, BF10 = 4.58) of the contextual modulations correlated between inverted eye matching and contrast detection tasks. Our results suggest that non-face-specialized high-level contextual mechanisms (inverted faces) work in connection to primary contextual mechanisms, but that the engagement of face-specialized mechanisms for upright faces obscures this connection. Such combined study of low- and high-level contextual modulations sheds new light on the functional relationship between different levels of the visual processing hierarchy, and thus on its functional organization.

Twenty years of investigation with the case of prosopagnosia PS to understand human face identity recognition. Part II: Neural basis

Patient PS sustained her dramatic brain injury in 1992, the same year as the first report of a neuroimaging study of human face recognition. The present paper complements the review on the functional nature of PS's prosopagnosia (part I), illustrating how her case study directly, i.e., through neuroimaging investigations of her brain structure and activity, but also indirectly, through neural studies performed on other clinical cases and neurotypical individuals, inspired and constrained neural models of human face recognition. In the dominant right hemisphere for face recognition in humans, PS's main lesion concerns (inputs to) the inferior occipital gyrus (IOG), in a region where face-selective activity is typically found in normal individuals ('Occipital Face Area', OFA). Her case study initially supported the criticality of this region for face identity recognition (FIR) and provided the impetus for transcranial magnetic stimulation (TMS), intracerebral electrical stimulation, and cortical surgery studies that have generally supported this view. Despite PS's right IOG lesion, typical face-selectivity is found anteriorly in the middle portion of the fusiform gyrus, a hominoid structure (termed the right 'Fusiform Face Area', FFA) that is widely considered to be the most important region for human face recognition. This finding led to the original proposal of direct anatomico-functional connections from early visual cortices to the FFA, bypassing the IOG/OFA (lulu), a hypothesis supported by further neuroimaging studies of PS, other neurological cases and neuro-typical individuals with original visual stimulation paradigms, data recordings and analyses. The proposal of a lack of sensitivity to face identity in PS's right FFA due to defective reentrant inputs from the IOG/FFA has also been supported by other cases, functional connectivity and cortical surgery studies. Overall, neural studies of, and based on, the case of prosopagnosia PS strongly question the hierarchical organization of the human neural face recognition system, supporting a more flexible and dynamic view of this key social brain function.

Face detection in untrained deep neural networks

Face-selective neurons are observed in the primate visual pathway and are considered as the basis of face detection in the brain. However, it has been debated as to whether this neuronal selectivity can arise innately or whether it requires training from visual experience. Here, using a hierarchical deep neural network model of the ventral visual stream, we suggest a mechanism in which face-selectivity arises in the complete absence of training. We found that units selective to faces emerge robustly in randomly initialized networks and that these units reproduce many characteristics observed in monkeys. This innate selectivity also enables the untrained network to perform face-detection tasks. Intriguingly, we observed that units selective to various non-face objects can also arise innately in untrained networks. Our results imply that the random feedforward connections in early, untrained deep neural networks may be sufficient for initializing primitive visual selectivity.

Rationalizing constraints on the capacity for cognitive control

Humans are remarkably limited in: (i) how many control-dependent tasks they can execute simultaneously, and (ii) how intensely they can focus on a single task. These limitations are universal assumptions of most theories of cognition. Yet, a rationale for why humans are subject to these constraints remains elusive. This feature review draws on recent insights from psychology, neuroscience, and machine learning, to suggest that constraints on cognitive control may result from a rational adaptation to fundamental, computational dilemmas in neural architectures. The reviewed literature implies that limitations in multitasking may result from a trade-off between learning efficacy and processing efficiency and that limitations in the intensity of commitment to a single task may reflect a trade-off between cognitive stability and flexibility.

Learning Invariant Object and Spatial View Representations in the Brain Using Slow Unsupervised Learning

First, neurophysiological evidence for the learning of invariant representations in the inferior temporal visual cortex is described. This includes object and face representations with invariance for position, size, lighting, view and morphological transforms in the temporal lobe visual cortex; global object motion in the cortex in the superior temporal sulcus; and spatial view representations in the hippocampus that are invariant with respect to eye position, head direction, and place. Second, computational mechanisms that enable the brain to learn these invariant representations are proposed. For the ventral visual system, one key adaptation is the use of information available in the statistics of the environment in slow unsupervised learning to learn transform-invariant representations of objects. This contrasts with deep supervised learning in artificial neural networks, which uses training with thousands of exemplars forced into different categories by neuronal teachers. Similar slow learning principles apply to the learning of global object motion in the dorsal visual system leading to the cortex in the superior temporal sulcus. The learning rule that has been explored in VisNet is an associative rule with a short-term memory trace. The feed-forward architecture has four stages, with convergence from stage to stage. This type of slow learning is implemented in the brain in hierarchically organized competitive neuronal networks with convergence from stage to stage, with only 4-5 stages in the hierarchy. Slow learning is also shown to help the learning of coordinate transforms using gain modulation in the dorsal visual system extending into the parietal cortex and retrosplenial cortex. Representations are learned that are in allocentric spatial view coordinates of locations in the world and that are independent of eye position, head direction, and the place where the individual is located. This enables hippocampal spatial view cells to use idiothetic, self-motion, signals for navigation when the view details are obscured for short periods.

Intrusive memories as conditioned responses to trauma cues: An empirically supported concept?

Intrusions in posttraumatic stress disorder (PTSD) are clinically understood as conditioned responses (CRs) to trauma-cues; however, experimental evidence for this is limited. We subjected 84 healthy participants to a differential conditioned-intrusion paradigm, where neutral faces served as conditioned stimuli (CSs) and aversive film clips as unconditioned stimuli (USs). While one group only completed acquisition, another group additionally received extinction. Subsequently, participants provided detailed e-diary intrusion reports. Several key findings emerged: First, participants in both groups re-experienced not only USs but also CSs as content of their intrusions. Second, intrusions were elicited by cues resembling CSs, USs, and experimental context. Third, extinction reduced probability and severity of US intrusions, and accelerated their decay, and this was particularly the case in participants showing greater cognitive (US-expectancy) and physiological (SCR) differential responding to CS+ vs. CS- at end of acquisition (i.e., conditionability). Similarly, extinction reduced CS-intrusion probability and severity, but only in participants with greater cognitive conditionability. These results support conditioning's role in re-experiencing in two critical ways: (1) Conditioning during trauma provides cues that not only function as reminder cues, but also as content of intrusions; (2) After strong conditioning, weakening the original CS-US relationship via extinction reduces intrusion formation after analogue-trauma.

Dysfunction of Magnocellular/dorsal Processing Stream in Schizophrenia

Background:

Patients with schizophrenia show not only cognitive, but also perceptual deficits. Perceptual deficits may affect different sensory modalities. Among these, the impairment of visual information processing is of particular relevance as demonstrated by the high incidence of visual disturbances. In recent years, the study of neurophysiological mechanisms that underlie visuo-perceptual, -spatial and -motor disorders in schizophrenia has increasingly attracted the interest of researchers.

Objective:

The study aims to review the existent literature on magnocellular/dorsal (occipitoparietal) visual processing stream impairment in schizophrenia. The impairment of relatively early stages of visual information processing was examined using experimental paradigms such as backward masking, contrast sensitivity, contour detection, and perceptual closure. The deficits of late processing stages were detected by examining visuo-spatial and -motor abilities.

Results:

Neurophysiological and behavioral studies support the existence of deficits in the processing of visual information along the magnocellular/dorsal pathway. These deficits appear to affect both early and late stages of visual information processing.

Conclusion:

The existence of disturbances in the early processing of visual information along the magnocellular/dorsal pathway is strongly supported by neurophysiological and behavioral observations. Early magnocellular dysfunction may provide a substrate for late dorsal processing impairment as well as higher-level cognition deficits.

Damasio's error - Prosopagnosia with intact within-category object recognition

The sudden inability to recognize individual faces following brain damage was first reported in a scientific journal 150 years ago and termed 'prosopagnosia' 70 years ago. While the term originally identified a face-selective neurological condition, it is now obscured by a sequence of imprecisions. First, prosopagnosia is routinely used to define symptoms of individual face recognition (IFR) difficulties in the context of visual object agnosia or other neurological conditions, or even in the normal population. Second, this over-expansive definition has lent support to a long-standing within-category recognition account of prosopagnosia, that is, that the impairment of IFR reflects a general impairment in recognizing within-category objects. However, stringent experimental studies of classical cases of prosopagnosia following brain damage show that their core impairment is not in recognizing physically similar exemplars within non-face object categories. Instead, the impairment presents specifically for recognizing exemplars of the category of faces. Moreover, compared to typical observers, the impairment appears even more severe for recognizing individual faces against physically dissimilar than similar distractors. Here, I argue that we need to limit accordingly our definition of prosopagnosia to a clinical (i.e., neurological) condition in which there is no basic-level object recognition impairment. Other criteria for prosopagnosia are proposed, with the hope that this conservative definition enables the study of human IFR processes in isolation, and supports progress in understanding the nature of these processes.

Reduced Performance During a Sentence Repetition Task by Continuous Theta-Burst Magnetic Stimulation of the Pre-supplementary Motor Area

The pre-supplementary motor area (pre-SMA) is engaged in speech comprehension under difficult circumstances such as poor acoustic signal quality or time-critical conditions. Previous studies found that left pre-SMA is activated when subjects listen to accelerated speech. Here, the functional role of pre-SMA was tested for accelerated speech comprehension by inducing a transient “virtual lesion” using continuous theta-burst stimulation (cTBS). Participants were tested (1) prior to (pre-baseline), (2) 10 min after (test condition for the cTBS effect), and (3) 60 min after stimulation (post-baseline) using a sentence repetition task (formant-synthesized at rates of 8, 10, 12, 14, and 16 syllables/s). Speech comprehension was quantified by the percentage of correctly reproduced speech material. For high speech rates, subjects showed decreased performance after cTBS of pre-SMA. Regarding the error pattern, the number of incorrect words without any semantic or phonological similarity to the target context increased, while related words decreased. Thus, the transient impairment of pre-SMA seems to affect its inhibitory function that normally eliminates erroneous speech material prior to speaking or, in case of perception, prior to encoding into a semantically/pragmatically meaningful message.

Functional Preference for Object Sounds and Voices in the Brain of Early Blind and Sighted Individuals

Sounds activate occipital regions in early blind individuals. However, how different sound categories map onto specific regions of the occipital cortex remains a matter of debate. We used fMRI to characterize brain responses of early blind and sighted individuals to familiar object sounds, human voices, and their respective low-level control sounds. In addition, sighted participants were tested while viewing pictures of faces, objects, and phase-scrambled control pictures. In both early blind and sighted, a double dissociation was evidenced in bilateral auditory cortices between responses to voices and object sounds: Voices elicited categorical responses in bilateral superior temporal sulci, whereas object sounds elicited categorical responses along the lateral fissure bilaterally, including the primary auditory cortex and planum temporale. Outside the auditory regions, object sounds also elicited categorical responses in the left lateral and in the ventral occipitotemporal regions in both groups. These regions also showed response preference for images of objects in the sighted group, thus suggesting a functional specialization that is independent of sensory input and visual experience. Between-group comparisons revealed that, only in the blind group, categorical responses to object sounds extended more posteriorly into the occipital cortex. Functional connectivity analyses evidenced a selective increase in the functional coupling between these reorganized regions and regions of the ventral occipitotemporal cortex in the blind group. In contrast, vocal sounds did not elicit preferential responses in the occipital cortex in either group. Nevertheless, enhanced voice-selective connectivity between the left temporal voice area and the right fusiform gyrus were found in the blind group. Altogether, these findings suggest that, in the absence of developmental vision, separate auditory categories are not equipotent in driving selective auditory recruitment of occipitotemporal regions and highlight the presence of domain-selective constraints on the expression of cross-modal plasticity.

Fixations Gate Species-Specific Responses to Free Viewing of Faces in the Human and Macaque Amygdala

Neurons in the primate amygdala respond prominently to faces. This implicates the amygdala in the processing of socially significant stimuli, yet its contribution to social perception remains poorly understood. We evaluated the representation of faces in the primate amygdala during naturalistic conditions by recording from both human and macaque amygdala neurons during free viewing of identical arrays of images with concurrent eye tracking. Neurons responded to faces only when they were fixated, suggesting that neuronal activity was gated by visual attention. Further experiments in humans utilizing covert attention confirmed this hypothesis. In both species, the majority of face-selective neurons preferred faces of conspecifics, a bias also seen behaviorally in first fixation preferences. Response latencies, relative to fixation onset, were shortest for conspecific-selective neurons and were ∼100 ms shorter in monkeys compared to humans. This argues that attention to faces gates amygdala responses, which in turn prioritize species-typical information for further processing.

Working memory, attention, and salience in active inference

The psychological concepts of working memory and attention are widely used in the cognitive and neuroscientific literatures. Perhaps because of the interdisciplinary appeal of these concepts, the same terms are often used to mean very different things. Drawing on recent advances in theoretical neurobiology, this paper tries to highlight the correspondence between these established psychological constructs and the formal processes implicit in mathematical descriptions of brain function. Here, we consider attention and salience from the perspective offered by active inference. Using variational principles and simulations, we use active inference to demonstrate how attention and salience can be disambiguated in terms of message passing between populations of neurons in cortical and subcortical structures. In brief, we suggest that salience is something that is afforded to actions that realise epistemic affordance, while attention per se is afforded to precise sensory evidence - or beliefs about the causes of sensations.

Rapid Processing of a Global Feature in the ON Visual Pathways of Behaving Monkeys

Visual objects are recognized by their features. Whereas, some features are based on simple components (i.e., local features, such as orientation of line segments), some features are based on the whole object (i.e., global features, such as an object having a hole in it). Over the past five decades, behavioral, physiological, anatomical, and computational studies have established a general model of vision, which starts from extracting local features in the lower visual pathways followed by a feature integration process that extracts global features in the higher visual pathways. This local-to-global model is successful in providing a unified account for a vast sets of perception experiments, but it fails to account for a set of experiments showing human visual systems' superior sensitivity to global features. Understanding the neural mechanisms underlying the "global-first" process will offer critical insights into new models of vision. The goal of the present study was to establish a non-human primate model of rapid processing of global features for elucidating the neural mechanisms underlying differential processing of global and local features. Monkeys were trained to make a saccade to a target in the black background, which was different from the distractors (white circle) in color (e.g., red circle target), local features (e.g., white square target), a global feature (e.g., white ring with a hole target) or their combinations (e.g., red square target). Contrary to the predictions of the prevailing local-to-global model, we found that (1) detecting a distinction or a change in the global feature was faster than detecting a distinction or a change in color or local features; (2) detecting a distinction in color was facilitated by a distinction in the global feature, but not in the local features; and (3) detecting the hole was interfered by the local features of the hole (e.g., white ring with a squared hole). These results suggest that monkey ON visual systems have a subsystem that is more sensitive to distinctions in the global feature than local features. They also provide the behavioral constraints for identifying the underlying neural substrates.

What can neuronal populations tell us about cognition?

Nowadays, it is possible to record the activity of hundreds of cells at the same time in behaving animals. However, these data are often treated and analyzed as if they consisted of many independently recorded neurons. How can neuronal populations be uniquely used to learn about cognition? We describe recent work that shows that populations of simultaneously recorded neurons are fundamental to understand the basis of decision-making, including processes such as ongoing deliberations and decision confidence, which generally fall outside the reach of single-cell analysis. Thus, neuronal population data allow addressing novel questions, but they also come with so far unsolved challenges.

New perspectives on the neurophysiology of primate amygdala emerging from the study of naturalistic social behaviors

A major challenge of primate neurophysiology, particularly in the domain of social neuroscience, is to adopt more natural behaviors without compromising the ability to relate patterns of neural activity to specific actions or sensory inputs. Traditional approaches have identified neural activity patterns in the amygdala in response to simplified versions of social stimuli such as static images of faces. As a departure from this reduced approach, single images of faces were replaced with arrays of images or videos of conspecifics. These stimuli elicited more natural behaviors and new types of neural responses: (1) attention-gated responses to faces, (2) selective responses to eye contact, and (3) selective responses to touch and somatosensory feedback during the production of facial expressions. An additional advance toward more natural social behaviors in the laboratory was the implementation of dyadic social interactions. Under these conditions, neurons encoded similarly rewards that monkeys delivered to self and to their social partner. These findings reinforce the value of bringing natural, ethologically valid, behavioral tasks under neurophysiological scrutiny. WIREs Cogn Sci 2018, 9:e1449. doi: 10.1002/wcs.1449 This article is categorized under: Psychology > Emotion and Motivation Neuroscience > Cognition Neuroscience > Physiology.

An emergent functional parcellation of the temporal cortex

The temporal lobe has been associated with various cognitive functions which include memory, auditory cognition and semantics. However, at a higher level of conceptualisation, all of the functions associated with the temporal lobe can be considered as lying along one major axis; from modality-specific to modality-general processing. This paper used a spectral reordering technique on resting-state and task-based functional data to extract the major organisational axis of the temporal lobe in a bottom-up, data-driven fashion. Independent parcellations were performed on resting-state scans from 71 participants and active semantic task scans from 23 participants acquired using dual echo gradient echo planar imaging in order to preserve signal in inferior temporal cortex. The resulting organisational axis was consistent (over dataset and hemisphere) and progressed from superior temporal gyrus and posterior inferior temporal cortex to ventrolateral anterior temporal cortex. A hard parcellation separated a posterior (superior temporal and posterior fusiform and inferior temporal gyri) and an anterior cluster (ventrolateral anterior temporal lobe). The functional connectivity of the hard clusters supported the hypothesis that the connectivity gradient separated modality-specific and modality-general regions. This hypothesis was then directly tested by performing a VOI analysis upon an independent semantic task-based data set including auditory and visually presented stimuli. This confirmed that the ventrolateral anterior aspects of the temporal lobe are associated with modality-general processes whilst posterior and superior aspects are specific to certain modalities, with the posterior inferior subregions involved in visual processes and superior regions involved in audition.

The structural and functional correlates of the efficiency in fearful face detection

Human visual system is found to be much efficient in searching for a fearful face. Some individuals are more sensitive to this threat-related stimulus. However, we still know little about the neural correlates of such variability. In the current study, we exploited a visual search paradigm, and asked the subjects to search for a fearful face or a target gender. Every subject showed a shallower search function for fearful face search than face gender search, indicating a stable fearful face advantage. We then used voxel-based morphometry (VBM) analysis and correlated this advantage to the gray matter volume (GMV) of some presumably face related cortical areas. The result revealed that only the left fusiform gyrus showed a significant positive correlation. Next, we defined the left fusiform gyrus as the seed region and calculated its resting state functional connectivity to the whole brain. Correlations were also calculated between fearful face advantage and these connectivities. In this analysis, we found positive correlations in the inferior parietal lobe and the ventral medial prefrontal cortex. These results suggested that the anatomical structure of the left fusiform gyrus might determine the search efficiency of fearful face, and frontoparietal attention network involved in this process through top-down attentional modulation.

The Theory of Localist Representation and of a Purely Abstract Cognitive System: The Evidence from Cortical Columns, Category Cells, and Multisensory Neurons

The debate about representation in the brain and the nature of the cognitive system has been going on for decades now. This paper examines the neurophysiological evidence, primarily from single cell recordings, to get a better perspective on both the issues. After an initial review of some basic concepts, the paper reviews the data from single cell recordings - in cortical columns and of category-selective and multisensory neurons. In neuroscience, columns in the neocortex (cortical columns) are understood to be a basic functional/computational unit. The paper reviews the fundamental discoveries about the columnar organization and finds that it reveals a massively parallel search mechanism. This columnar organization could be the most extensive neurophysiological evidence for the widespread use of localist representation in the brain. The paper also reviews studies of category-selective cells. The evidence for category-selective cells reveals that localist representation is also used to encode complex abstract concepts at the highest levels of processing in the brain. A third major issue is the nature of the cognitive system in the brain and whether there is a form that is purely abstract and encoded by single cells. To provide evidence for a single-cell based purely abstract cognitive system, the paper reviews some of the findings related to multisensory cells. It appears that there is widespread usage of multisensory cells in the brain in the same areas where sensory processing takes place. Plus there is evidence for abstract modality invariant cells at higher levels of cortical processing. Overall, that reveals the existence of a purely abstract cognitive system in the brain. The paper also argues that since there is no evidence for dense distributed representation and since sparse representation is actually used to encode memories, there is actually no evidence for distributed representation in the brain. Overall, it appears that, at an abstract level, the brain is a massively parallel, distributed computing system that is symbolic. The paper also explains how grounded cognition and other theories of the brain are fully compatible with localist representation and a purely abstract cognitive system.

A neural model of normal and abnormal learning and memory consolidation: adaptively timed conditioning, hippocampus, amnesia, neurotrophins, and consciousness

How do the hippocampus and amygdala interact with thalamocortical systems to regulate cognitive and cognitive-emotional learning? Why do lesions of thalamus, amygdala, hippocampus, and cortex have differential effects depending on the phase of learning when they occur? In particular, why is the hippocampus typically needed for trace conditioning, but not delay conditioning, and what do the exceptions reveal? Why do amygdala lesions made before or immediately after training decelerate conditioning while those made later do not? Why do thalamic or sensory cortical lesions degrade trace conditioning more than delay conditioning? Why do hippocampal lesions during trace conditioning experiments degrade recent but not temporally remote learning? Why do orbitofrontal cortical lesions degrade temporally remote but not recent or post-lesion learning? How is temporally graded amnesia caused by ablation of prefrontal cortex after memory consolidation? How are attention and consciousness linked during conditioning? How do neurotrophins, notably brain-derived neurotrophic factor (BDNF), influence memory formation and consolidation? Is there a common output path for learned performance? A neural model proposes a unified answer to these questions that overcome problems of alternative memory models.

A face-selective ventral occipito-temporal map of the human brain with intracerebral potentials

Human neuroimaging studies have identified a network of distinct face-selective regions in the ventral occipito-temporal cortex (VOTC), with a right hemispheric dominance. To date, there is no evidence for this hemispheric and regional specialization with direct measures of brain activity. To address this gap in knowledge, we recorded local neurophysiological activity from 1,678 contact electrodes implanted in the VOTC of a large group of epileptic patients (n = 28). They were presented with natural images of objects at a rapid fixed rate (six images per second: 6 Hz), with faces interleaved as every fifth stimulus (i.e., 1.2 Hz). High signal-to-noise ratio face-selective responses were objectively (i.e., exactly at the face stimulation frequency) identified and quantified throughout the whole VOTC. Face-selective responses were widely distributed across the whole VOTC, but also spatially clustered in specific regions. Among these regions, the lateral section of the right middle fusiform gyrus showed the largest face-selective response by far, offering, to our knowledge, the first supporting evidence of two decades of neuroimaging observations with direct neural measures. In addition, three distinct regions with a high proportion of face-selective responses were disclosed in the right ventral anterior temporal lobe, a region that is undersampled in neuroimaging because of magnetic susceptibility artifacts. A high proportion of contacts responding only to faces (i.e., "face-exclusive" responses) were found in these regions, suggesting that they contain populations of neurons involved in dedicated face-processing functions. Overall, these observations provide a comprehensive mapping of visual category selectivity in the whole human VOTC with direct neural measures.

On the domain-specificity of the visual and non-visual face-selective regions

What happens in our brains when we see a face? The neural mechanisms of face processing - namely, the face-selective regions - have been extensively explored. Research has traditionally focused on visual cortex face-regions; more recently, the role of face-regions outside the visual cortex (i.e., non-visual-cortex face-regions) has been acknowledged as well. The major quest today is to reveal the functional role of each this region in face processing. To make progress in this direction, it is essential to understand the extent to which the face-regions, and particularly the non-visual-cortex face-regions, process only faces (i.e., face-specific, domain-specific processing) or rather are involved in a more domain-general cognitive processing. In the current functional MRI study, we systematically examined the activity of the whole face-network during face-unrelated reading task (i.e., written meaningful sentences with content unrelated to faces/people and non-words). We found that the non-visual-cortex (i.e., right lateral prefrontal cortex and posterior superior temporal sulcus), but not the visual cortex face-regions, responded significantly stronger to sentences than to non-words. In general, some degree of sentence selectivity was found in all non-visual-cortex cortex. Present result highlights the possibility that the processing in the non-visual-cortex face-selective regions might not be exclusively face-specific, but rather more or even fully domain-general. In this paper, we illustrate how the knowledge about domain-general processing in face-regions can help to advance our general understanding of face processing mechanisms. Our results therefore suggest that the problem of face processing should be approached in the broader scope of cognition in general.

Demonstrating the Acquired Familiarity of Faces by Using a Gender-Decision Task

We report an experiment in which participants made gender judgments (male or female) to faces. There were three groups of faces: unfamiliar, familiar (celebrities), and a set which had been learned earlier by the participants during the experimental session. The principal purpose of this study was to establish an indirect measure of assessing whether faces have become familiar through learning that does not require overt recognition. Bruce and Young's (1986 British Journal of Psychology77 305–327) model of face recognition suggests face-processing tasks are independent of one another and so familiarity should have no impact on the time taken to perform gender decisions. However, recent studies have suggested that some face processes are not completely independent. A gender judgment is a simple task which would be useful in face-learning experiments. We examined whether exposure to previously novel faces facilitates a later gender decision to those faces. During a learning stage, participants viewed a set of unfamiliar faces. At test, participants were able to assign gender faster to previously familiar (famous) faces and learned faces than they were to unfamiliar faces. Therefore familiarity can influence the speed at which gender is analysed. We explain our findings with reference to the Burton et al (1990 British Journal of Psychology81 361–380) interactive activation and competition (IAC) model of face recognition and discuss how the gender judgment might be employed as a means of tracking the acquisition of familiarity in face-learning studies.

A computational exploration of complementary learning mechanisms in the primate ventral visual pathway

In order to develop transformation invariant representations of objects, the visual system must make use of constraints placed upon object transformation by the environment. For example, objects transform continuously from one point to another in both space and time. These two constraints have been exploited separately in order to develop translation and view invariance in a hierarchical multilayer model of the primate ventral visual pathway in the form of continuous transformation learning and temporal trace learning. We show for the first time that these two learning rules can work cooperatively in the model. Using these two learning rules together can support the development of invariance in cells and help maintain object selectivity when stimuli are presented over a large number of locations or when trained separately over a large number of viewing angles.

Dynamic Functional Brain Connectivity for Face Perception

Face perception is mediated by a distributed brain network comprised of the core system at occipito-temporal areas and the extended system at other relevant brain areas involving bilateral hemispheres. In this study we explored how the brain connectivity changes over the time for face-sensitive processing. We investigated the dynamic functional connectivity in face perception by analyzing time-dependent EEG phase synchronization in four different frequency bands: theta (4-7 Hz), alpha (8-14 Hz), beta (15-24 Hz), and gamma (25-45 Hz) bands in the early stages of face processing from 30 to 300 ms. High-density EEG were recorded from subjects who were passively viewing faces, buildings, and chairs. The dynamic connectivity within the core system and between the extended system were investigated. Significant differences between faces and non-faces mainly appear in theta band connectivity: (1) at the time segment of 90-120 ms between parietal area and occipito-temporal area in the right hemisphere, and (2) at the time segment of 150-180 ms between bilateral occipito-temporal areas. These results indicate (1) the importance of theta-band connectivity in the face-sensitive processing, and (2) that different parts of network are involved for the initial stage of face categorization and the stage of face structural encoding.

The impact of parent-child interaction on brain structures: cross-sectional and longitudinal analyses

There is a vast amount of evidence from psychological studies that the amount of parent-child interaction affects the development of children's verbal skills and knowledge. However, despite the vast amount of literature, brain structural development associated with the amount of parent-child interaction has never been investigated. In the present human study, we used voxel-based morphometry to measure regional gray matter density (rGMD) and examined cross-sectional correlations between the amount of time spent with parents and rGMD among 127 boys and 135 girls. We also assessed correlations between the amount of time spent with parents and longitudinal changes that occurred a few years later among 106 boys and 102 girls. After correcting for confounding factors, we found negative effects of spending time with parents on rGMD in areas in the bilateral superior temporal gyrus (STG) via cross-sectional analyses as well as in the contingent areas of the right STG. We also confirmed positive effects of spending time with parents on the Verbal Comprehension score in cross-sectional and longitudinal analyses. rGMD in partly overlapping or contingent areas of the right STG was negatively correlated with age and the Verbal Comprehension score in cross-sectional analyses. Subsequent analyses revealed verbal parent-child interactions have similar effects on Verbal Comprehension scores and rGMD in the right STG in both cross-sectional and longitudinal analyses. These findings indicate that parent-child interactions affect the right STG, which may be associated with verbal skills.

The effect of face inversion for neurons inside and outside fMRI-defined face-selective cortical regions

It is widely believed that face processing in the primate brain occurs in a network of category-selective cortical regions. Combined functional MRI (fMRI)-single-cell recording studies in macaques have identified high concentrations of neurons that respond more to faces than objects within face-selective patches. However, cells with a preference for faces over objects are also found scattered throughout inferior temporal (IT) cortex, raising the question whether face-selective cells inside and outside of the face patches differ functionally. Here, we compare the properties of face-selective cells inside and outside of face-selective patches in the IT cortex by means of an image manipulation that reliably disrupts behavior toward face processing: inversion. We recorded IT neurons from two fMRI-defined face-patches (ML and AL) and a region outside of the face patches (herein labeled OUT) during upright and inverted face stimulation. Overall, turning faces upside down reduced the firing rate of face-selective cells. However, there were differences among the recording regions. First, the reduced neuronal response for inverted faces was independent of stimulus position, relative to fixation, in the face-selective patches (ML and AL) only. Additionally, the effect of inversion for face-selective cells in ML, but not those in AL or OUT, was impervious to whether the neurons were initially searched for using upright or inverted stimuli. Collectively, these results show that face-selective cells differ in their functional characteristics depending on their anatomicofunctional location, suggesting that upright faces are preferably coded by face-selective cells inside but not outside of the fMRI-defined face-selective regions of the posterior IT cortex.

Face recognition in primates: a cross-species study

Recognition for human faces, monkey faces, and objects was assessed in both adult humans (Homo sapiens) and monkeys (Macaca mulatta) with a visual paired-comparison task. The results demonstrated that while both species showed strong novelty preference for objects, human participants showed novelty preference for human faces but not for monkey faces, and vice versa for the monkeys. This `species-specific effect' in face recognition is discussed in relation with data on both the `other-race effect' observed in humans and the effects of experience or training on face recognition processes in primates.

A feed-forward spiking model of shape-coding by IT cells

The ability to recognize a shape is linked to figure-ground (FG) organization. Cell preferences appear to be correlated across contrast-polarity reversals and mirror reversals of polygon displays, but not so much across FG reversals. Here we present a network structure which explains both shape-coding by simulated IT cells and suppression of responses to FG reversed stimuli. In our model FG segregation is achieved before shape discrimination, which is itself evidenced by the difference in spiking onsets of a pair of output cells. The studied example also includes feature extraction and illustrates a classification of binary images depending on the dominance of vertical or horizontal borders.

Face recognition increases during saccade preparation

Face perception is integral to human perception system as it underlies social interactions. Saccadic eye movements are frequently made to bring interesting visual information, such as faces, onto the fovea for detailed processing. Just before eye movement onset, the processing of some basic features, such as the orientation, of an object improves at the saccade landing point. Interestingly, there is also evidence that indicates faces are processed in early visual processing stages similar to basic features. However, it is not known whether this early enhancement of processing includes face recognition. In this study, three experiments were performed to map the timing of face presentation to the beginning of the eye movement in order to evaluate pre-saccadic face recognition. Faces were found to be similarly processed as simple objects immediately prior to saccadic movements. Starting ∼ 120 ms before a saccade to a target face, independent of whether or not the face was surrounded by other faces, the face recognition gradually improved and the critical spacing of the crowding decreased as saccade onset was approaching. These results suggest that an upcoming saccade prepares the visual system for new information about faces at the saccade landing site and may reduce the background in a crowd to target the intended face. This indicates an important role of pre-saccadic eye movement signals in human face recognition.

Robust selectivity for faces in the human amygdala in the absence of expressions

There is a well-established posterior network of cortical regions that plays a central role in face processing and that has been investigated extensively. In contrast, although responsive to faces, the amygdala is not considered a core face-selective region, and its face selectivity has never been a topic of systematic research in human neuroimaging studies. Here, we conducted a large-scale group analysis of fMRI data from 215 participants. We replicated the posterior network observed in prior studies but found equally robust and reliable responses to faces in the amygdala. These responses were detectable in most individual participants, but they were also highly sensitive to the initial statistical threshold and habituated more rapidly than the responses in posterior face-selective regions. A multivariate analysis showed that the pattern of responses to faces across voxels in the amygdala had high reliability over time. Finally, functional connectivity analyses showed stronger coupling between the amygdala and posterior face-selective regions during the perception of faces than during the perception of control visual categories. These findings suggest that the amygdala should be considered a core face-selective region.

Toward a unified model of face and object recognition in the human visual system

Our understanding of the mechanisms and neural substrates underlying visual recognition has made considerable progress over the past 30 years. During this period, accumulating evidence has led many scientists to conclude that objects and faces are recognised in fundamentally distinct ways, and in fundamentally distinct cortical areas. In the psychological literature, in particular, this dissociation has led to a palpable disconnect between theories of how we process and represent the two classes of object. This paper follows a trend in part of the recognition literature to try to reconcile what we know about these two forms of recognition by considering the effects of learning. Taking a widely accepted, self-organizing model of object recognition, this paper explains how such a system is affected by repeated exposure to specific stimulus classes. In so doing, it explains how many aspects of recognition generally regarded as unusual to faces (holistic processing, configural processing, sensitivity to inversion, the other-race effect, the prototype effect, etc.) are emergent properties of category-specific learning within such a system. Overall, the paper describes how a single model of recognition learning can and does produce the seemingly very different types of representation associated with faces and objects.

How lateral connections and spiking dynamics may separate multiple objects moving together

Over successive stages, the ventral visual system of the primate brain develops neurons that respond selectively to particular objects or faces with translation, size and view invariance. The powerful neural representations found in Inferotemporal cortex form a remarkably rapid and robust basis for object recognition which belies the difficulties faced by the system when learning in natural visual environments. A central issue in understanding the process of biological object recognition is how these neurons learn to form separate representations of objects from complex visual scenes composed of multiple objects. We show how a one-layer competitive network comprised of ‘spiking’ neurons is able to learn separate transformation-invariant representations (exemplified by one-dimensional translations) of visual objects that are always seen together moving in lock-step, but separated in space. This is achieved by combining ‘Mexican hat’ functional lateral connectivity with cell firing-rate adaptation to temporally segment input representations of competing stimuli through anti-phase oscillations (perceptual cycles). These spiking dynamics are quickly and reliably generated, enabling selective modification of the feed-forward connections to neurons in the next layer through Spike-Time-Dependent Plasticity (STDP), resulting in separate translation-invariant representations of each stimulus. Variations in key properties of the model are investigated with respect to the network’s ability to develop appropriate input representations and subsequently output representations through STDP. Contrary to earlier rate-coded models of this learning process, this work shows how spiking neural networks may learn about more than one stimulus together without suffering from the ‘superposition catastrophe’. We take these results to suggest that spiking dynamics are key to understanding biological visual object recognition.

Neuroethology of primate social behavior

A neuroethological approach to human and nonhuman primate behavior and cognition predicts biological specializations for social life. Evidence reviewed here indicates that ancestral mechanisms are often duplicated, repurposed, and differentially regulated to support social behavior. Focusing on recent research from nonhuman primates, we describe how the primate brain might implement social functions by coopting and extending preexisting mechanisms that previously supported nonsocial functions. This approach reveals that highly specialized mechanisms have evolved to decipher the immediate social context, and parallel circuits have evolved to translate social perceptual signals and nonsocial perceptual signals into partially integrated social and nonsocial motivational signals, which together inform general-purpose mechanisms that command behavior. Differences in social behavior between species, as well as between individuals within a species, result in part from neuromodulatory regulation of these neural circuits, which itself appears to be under partial genetic control. Ultimately, intraspecific variation in social behavior has differential fitness consequences, providing fundamental building blocks of natural selection. Our review suggests that the neuroethological approach to primate behavior may provide unique insights into human psychopathology.

Bayesian computation emerges in generic cortical microcircuits through spike-timing-dependent plasticity

The principles by which networks of neurons compute, and how spike-timing dependent plasticity (STDP) of synaptic weights generates and maintains their computational function, are unknown. Preceding work has shown that soft winner-take-all (WTA) circuits, where pyramidal neurons inhibit each other via interneurons, are a common motif of cortical microcircuits. We show through theoretical analysis and computer simulations that Bayesian computation is induced in these network motifs through STDP in combination with activity-dependent changes in the excitability of neurons. The fundamental components of this emergent Bayesian computation are priors that result from adaptation of neuronal excitability and implicit generative models for hidden causes that are created in the synaptic weights through STDP. In fact, a surprising result is that STDP is able to approximate a powerful principle for fitting such implicit generative models to high-dimensional spike inputs: Expectation Maximization. Our results suggest that the experimentally observed spontaneous activity and trial-to-trial variability of cortical neurons are essential features of their information processing capability, since their functional role is to represent probability distributions rather than static neural codes. Furthermore it suggests networks of Bayesian computation modules as a new model for distributed information processing in the cortex.

Recognizing sights, smells, and sounds with gnostic fields

Mammals rely on vision, audition, and olfaction to remotely sense stimuli in their environment. Determining how the mammalian brain uses this sensory information to recognize objects has been one of the major goals of psychology and neuroscience. Likewise, researchers in computer vision, machine audition, and machine olfaction have endeavored to discover good algorithms for stimulus classification. Almost 50 years ago, the neuroscientist Jerzy Konorski proposed a theoretical model in his final monograph in which competing sets of “gnostic” neurons sitting atop sensory processing hierarchies enabled stimuli to be robustly categorized, despite variations in their presentation. Much of what Konorski hypothesized has been remarkably accurate, and neurons with gnostic-like properties have been discovered in visual, aural, and olfactory brain regions. Surprisingly, there have not been any attempts to directly transform his theoretical model into a computational one. Here, I describe the first computational implementation of Konorski's theory. The model is not domain specific, and it surpasses the best machine learning algorithms on challenging image, music, and olfactory classification tasks, while also being simpler. My results suggest that criticisms of exemplar-based models of object recognition as being computationally intractable due to limited neural resources are unfounded.