Modeling naturalistic face processing in humans with deep convolutional neural networks.

Deep convolutional neural networks (DCNNs) trained for face identification can rival and even exceed human-level performance. The ways in which the internal face representations in DCNNs relate to human cognitive representations and brain activity are not well understood. Nearly all previous studies focused on static face image processing with rapid display times and ignored the processing of naturalistic, dynamic information. To address this gap, we developed the largest naturalistic dynamic face stimulus set in human neuroimaging research (700+ naturalistic video clips of unfamiliar faces). We used this naturalistic dataset to compare representational geometries estimated from DCNNs, behavioral responses, and brain responses. We found that DCNN representational geometries were consistent across architectures, cognitive representational geometries were consistent across raters in a behavioral arrangement task, and neural representational geometries in face areas were consistent across brains. Representational geometries in late, fully connected DCNN layers, which are optimized for individuation, were much more weakly correlated with cognitive and neural geometries than were geometries in late-intermediate layers. The late-intermediate face-DCNN layers successfully matched cognitive representational geometries, as measured with a behavioral arrangement task that primarily reflected categorical attributes, and correlated with neural representational geometries in known face-selective topographies. Our study suggests that current DCNNs successfully capture neural cognitive processes for categorical attributes of faces but less accurately capture individuation and dynamic features.

Shared neural codes for visual and semantic information about familiar faces in a common representational space.

Processes evoked by seeing a personally familiar face encompass recognition of visual appearance and activation of social and person knowledge. Whereas visual appearance is the same for all viewers, social and person knowledge may be more idiosyncratic. Using between-subject multivariate decoding of hyperaligned functional magnetic resonance imaging data, we investigated whether representations of personally familiar faces in different parts of the distributed neural system for face perception are shared across individuals who know the same people. We found that the identities of both personally familiar and merely visually familiar faces were decoded accurately across brains in the core system for visual processing, but only the identities of personally familiar faces could be decoded across brains in the extended system for processing nonvisual information associated with faces. Our results show that personal interactions with the same individuals lead to shared neural representations of both the seen and unseen features that distinguish their identities.

Hybrid hyperalignment: A single high-dimensional model of shared information embedded in cortical patterns of response and functional connectivity.

Shared information content is represented across brains in idiosyncratic functional topographies. Hyperalignment addresses these idiosyncrasies by using neural responses to project individuals' brain data into a common model space while maintaining the geometric relationships between distinct patterns of activity or connectivity. The dimensions of this common model capture functional profiles that are shared across individuals such as cortical response profiles collected during a common time-locked stimulus presentation (e.g. movie viewing) or functional connectivity profiles. Hyperalignment can use either response-based or connectivity-based input data to derive transformations that project individuals' neural data from anatomical space into the common model space. Previously, only response or connectivity profiles were used in the derivation of these transformations. In this study, we developed a new hyperalignment algorithm, hybrid hyperalignment, that derives transformations based on both response-based and connectivity-based information. We used three different movie-viewing fMRI datasets to test the performance of our new algorithm. Hybrid hyperalignment derives a single common model space that aligns response-based information as well as or better than response hyperalignment while simultaneously aligning connectivity-based information better than connectivity hyperalignment. These results suggest that a single common information space can encode both shared cortical response and functional connectivity profiles across individuals.

Naturalistic stimuli reveal a dominant role for agentic action in visual representation.

Naturalistic, dynamic movies evoke strong, consistent, and information-rich patterns of activity over a broad expanse of cortex and engage multiple perceptual and cognitive systems in parallel. The use of naturalistic stimuli enables functional brain imaging research to explore cognitive domains that are poorly sampled in highly-controlled experiments. These domains include perception and understanding of agentic action, which plays a larger role in visual representation than was appreciated from experiments using static, controlled stimuli.

Predicting individual face-selective topography using naturalistic stimuli.

Subject-specific, functionally defined areas are conventionally estimated with functional localizers and a simple contrast analysis between responses to different stimulus categories. Compared with functional localizers, naturalistic stimuli provide several advantages such as stronger and widespread brain activation, greater engagement, and increased subject compliance. In this study we demonstrate that a subject's idiosyncratic functional topography can be estimated with high fidelity from that subject's fMRI data obtained while watching a naturalistic movie using hyperalignment to project other subjects' localizer data into that subject's idiosyncratic cortical anatomy. These findings lay the foundation for developing an efficient tool for mapping functional topographies for a wide range of perceptual and cognitive functions in new subjects based only on fMRI data collected while watching an engaging, naturalistic stimulus and other subjects' localizer data from a normative sample.

How familiarity warps representation in the face space.

Recognition of familiar as compared to unfamiliar faces is robust and resistant to marked image distortion or degradation. Here we tested the flexibility of familiar face recognition with a morphing paradigm where the appearance of a personally familiar face was mixed with the appearance of a stranger (Experiment 1) and the appearance of one's own face with the appearance of a familiar face and the appearance of a stranger (Experiment 2). The aim of the two experiments was to assess how categorical boundaries for recognition of identity are affected by familiarity. We found a narrower categorical boundary for the identity of personally familiar faces when they were mixed with unfamiliar identities as compared to the control condition, in which the appearance of two unfamiliar faces was mixed. Our results suggest that familiarity warps the representational geometry of face space, amplifying perceptual distances for small changes in the appearance of familiar faces that are inconsistent with the structural features that define their identities.

Disentangling the Representation of Identity from Head View Along the Human Face Processing Pathway.

Neural models of a distributed system for face perception implicate a network of regions in the ventral visual stream for recognition of identity. Here, we report a functional magnetic resonance imaging (fMRI) neural decoding study in humans that shows that this pathway culminates in the right inferior frontal cortex face area (rIFFA) with a representation of individual identities that has been disentangled from variable visual features in different images of the same person. At earlier stages in the pathway, processing begins in early visual cortex and the occipital face area with representations of head view that are invariant across identities, and proceeds to an intermediate level of representation in the fusiform face area in which identity is emerging but still entangled with head view. Three-dimensional, view-invariant representation of identities in the rIFFA may be the critical link to the extended system for face perception, affording activation of person knowledge and emotional responses to familiar faces.

Attention Selectively Reshapes the Geometry of Distributed Semantic Representation.

Humans prioritize different semantic qualities of a complex stimulus depending on their behavioral goals. These semantic features are encoded in distributed neural populations, yet it is unclear how attention might operate across these distributed representations. To address this, we presented participants with naturalistic video clips of animals behaving in their natural environments while the participants attended to either behavior or taxonomy. We used models of representational geometry to investigate how attentional allocation affects the distributed neural representation of animal behavior and taxonomy. Attending to animal behavior transiently increased the discriminability of distributed population codes for observed actions in anterior intraparietal, pericentral, and ventral temporal cortices. Attending to animal taxonomy while viewing the same stimuli increased the discriminability of distributed animal category representations in ventral temporal cortex. For both tasks, attention selectively enhanced the discriminability of response patterns along behaviorally relevant dimensions. These findings suggest that behavioral goals alter how the brain extracts semantic features from the visual world. Attention effectively disentangles population responses for downstream read-out by sculpting representational geometry in late-stage perceptual areas.

How the Human Brain Represents Perceived Dangerousness or "Predacity" of Animals.

UNLABELLED: Common or folk knowledge about animals is dominated by three dimensions: (1) level of cognitive complexity or "animacy;" (2) dangerousness or "predacity;" and (3) size. We investigated the neural basis of the perceived dangerousness or aggressiveness of animals, which we refer to more generally as "perception of threat." Using functional magnetic resonance imaging (fMRI), we analyzed neural activity evoked by viewing images of animal categories that spanned the dissociable semantic dimensions of threat and taxonomic class. The results reveal a distributed network for perception of threat extending along the right superior temporal sulcus. We compared neural representational spaces with target representational spaces based on behavioral judgments and a computational model of early vision and found a processing pathway in which perceived threat emerges as a dominant dimension: whereas visual features predominate in early visual cortex and taxonomy in lateral occipital and ventral temporal cortices, these dimensions fall away progressively from posterior to anterior temporal cortices, leaving threat as the dominant explanatory variable. Our results suggest that the perception of threat in the human brain is associated with neural structures that underlie perception and cognition of social actions and intentions, suggesting a broader role for these regions than has been thought previously, one that includes the perception of potential threat from agents independent of their biological class. SIGNIFICANCE STATEMENT: For centuries, philosophers have wondered how the human mind organizes the world into meaningful categories and concepts. Today this question is at the core of cognitive science, but our focus has shifted to understanding how knowledge manifests in dynamic activity of neural systems in the human brain. This study advances the young field of empirical neuroepistemology by characterizing the neural systems engaged by an important dimension in our cognitive representation of the animal kingdom ontological subdomain: how the brain represents the perceived threat, dangerousness, or "predacity" of animals. Our findings reveal how activity for domain-specific knowledge of animals overlaps the social perception networks of the brain, suggesting domain-general mechanisms underlying the representation of conspecifics and other animals.

Processing of invisible social cues.

Successful interactions between people are dependent on rapid recognition of social cues. We investigated whether head direction--a powerful social signal--is processed in the absence of conscious awareness. We used continuous flash interocular suppression to render stimuli invisible and compared the reaction time for face detection when faces were turned towards the viewer and turned slightly away. We found that faces turned towards the viewer break through suppression faster than faces that are turned away, regardless of eye direction. Our results suggest that detection of a face with attention directed at the viewer occurs even in the absence of awareness of that face. While previous work has demonstrated that stimuli that signal threat are processed without awareness, our data suggest that the social relevance of a face, defined more broadly, is evaluated in the absence of awareness.

Cross-movie prediction of individualized functional topography.

Participant-specific, functionally defined brain areas are usually mapped with functional localizers and estimated by making contrasts between responses to single categories of input. Naturalistic stimuli engage multiple brain systems in parallel, provide more ecologically plausible estimates of real-world statistics, and are friendly to special populations. The current study shows that cortical functional topographies in individual participants can be estimated with high fidelity from naturalistic stimuli. Importantly, we demonstrate that robust, individualized estimates can be obtained even when participants watched different movies, were scanned with different parameters/scanners, and were sampled from different institutes across the world. Our results create a foundation for future studies that allow researchers to estimate a broad range of functional topographies based on naturalistic movies and a normative database, making it possible to integrate high-level cognitive functions across datasets from laboratories worldwide.

An fMRI dataset in response to "The Grand Budapest Hotel", a socially-rich, naturalistic movie.

Naturalistic stimuli evoke strong, consistent, and information-rich patterns of brain activity, and engage large extents of the human brain. They allow researchers to compare highly similar brain responses across subjects, and to study how complex representations are encoded in brain activity. Here, we describe and share a dataset where 25 subjects watched part of the feature film "The Grand Budapest Hotel" by Wes Anderson. The movie has a large cast with many famous actors. Throughout the story, the camera shots highlight faces and expressions, which are fundamental to understand the complex narrative of the movie. This movie was chosen to sample brain activity specifically related to social interactions and face processing. This dataset provides researchers with fMRI data that can be used to explore social cognitive processes and face processing, adding to the existing neuroimaging datasets that sample brain activity with naturalistic movies.

Idiosyncratic, Retinotopic Bias in Face Identification Modulated by Familiarity.

The perception of gender and age of unfamiliar faces is reported to vary idiosyncratically across retinal locations such that, for example, the same androgynous face may appear to be male at one location but female at another. Here, we test spatial heterogeneity for the recognition of the identity of personally familiar faces in human participants. We found idiosyncratic biases that were stable within participants and that varied more across locations for low as compared to high familiar faces. These data suggest that like face gender and age, face identity is processed, in part, by independent populations of neurons monitoring restricted spatial regions and that the recognition responses vary for the same face across these different locations. Moreover, repeated and varied social interactions appear to lead to adjustments of these independent face recognition neurons so that the same familiar face is eventually more likely to elicit the same recognition response across widely separated visual field locations. We provide a mechanistic account of this reduced retinotopic bias based on computational simulations.

Modeling Semantic Encoding in a Common Neural Representational Space.

Encoding models for mapping voxelwise semantic tuning are typically estimated separately for each individual, limiting their generalizability. In the current report, we develop a method for estimating semantic encoding models that generalize across individuals. Functional MRI was used to measure brain responses while participants freely viewed a naturalistic audiovisual movie. Word embeddings capturing agent-, action-, object-, and scene-related semantic content were assigned to each imaging volume based on an annotation of the film. We constructed both conventional within-subject semantic encoding models and between-subject models where the model was trained on a subset of participants and validated on a left-out participant. Between-subject models were trained using cortical surface-based anatomical normalization or surface-based whole-cortex hyperalignment. We used hyperalignment to project group data into an individual's unique anatomical space via a common representational space, thus leveraging a larger volume of data for out-of-sample prediction while preserving the individual's fine-grained functional-anatomical idiosyncrasies. Our findings demonstrate that anatomical normalization degrades the spatial specificity of between-subject encoding models relative to within-subject models. Hyperalignment, on the other hand, recovers the spatial specificity of semantic tuning lost during anatomical normalization, and yields model performance exceeding that of within-subject models.

Neural systems for recognition of familiar faces.

Immediate access to information about people that we encounter is an essential requirement for effective social interactions. In this manuscript we briefly review our work and work of others on familiar face recognition and propose a modified version of our model of neural systems for face perception with a special emphasis on processes associated with recognition of familiar faces. We argue that visual appearance is only one component of successful recognition of familiar individuals. Other fundamental aspects include the retrieval of "person knowledge" - the representation of the personal traits, intentions, and outlook of someone we know - and the emotional response we experience when seeing a familiar individual. Specifically, we hypothesize that the "theory of mind" areas, that have been implicated in social and cognitive functions other than face perception, play an essential role in the spontaneous activation of person knowledge associated with the recognition of familiar individuals. The amygdala and the insula, structures that are involved in the representation of emotion, also are part of the distributed network of areas that are modulated by familiarity, reflecting the role of emotion in face recognition.

Spontaneous retrieval of affective person knowledge in face perception.

In a functional magnetic resonance imaging experiment, we explored whether affective person knowledge based on memories formed from minimal information is spontaneously retrieved in face perception. In the first stage of the experiment, participants were presented with 120 unfamiliar faces. Each face was presented with a description of one of four types of behaviors: aggressive, disgusting, neutral, and nice. In the second stage, participants were scanned while engaged in a one-back recognition task in which they saw the faces that were associated with behaviors and 30 novel faces. Although this task is a simple perceptual task that neither demands person evaluation nor retrieval of person knowledge, neural responses to faces differed as a function of the behaviors. Faces associated with behaviors evoked stronger activity than did novel faces in regions implicated in social cognition--anterior paracingulate cortex and superior temporal sulcus. Explicit memory for the behaviors enhanced the neural response in these regions. Faces associated with disgusting behaviors evoked stronger activity in left anterior insula than did faces associated with aggressive behaviors. This effect was equally strong for faces associated with explicitly recalled behaviors and faces associated with non-recalled behaviors. The findings suggest that affective person knowledge acquired from minimal information is spontaneously retrieved in face perception, engaging neural systems for analysis of social cognition and emotions.

Reading Faces: From Features to Recognition.

Chang and Tsao recently reported that the monkey face patch system encodes facial identity in a space of facial features as opposed to exemplars. Here, we discuss how such coding might contribute to face recognition, emphasizing the critical role of learning and interactions with other brain areas for optimizing the recognition of familiar faces.

The neural representation of personally familiar and unfamiliar faces in the distributed system for face perception.

Personally familiar faces are processed more robustly and efficiently than unfamiliar faces. The human face processing system comprises a core system that analyzes the visual appearance of faces and an extended system for the retrieval of person-knowledge and other nonvisual information. We applied multivariate pattern analysis to fMRI data to investigate aspects of familiarity that are shared by all familiar identities and information that distinguishes specific face identities from each other. Both identity-independent familiarity information and face identity could be decoded in an overlapping set of areas in the core and extended systems. Representational similarity analysis revealed a clear distinction between the two systems and a subdivision of the core system into ventral, dorsal and anterior components. This study provides evidence that activity in the extended system carries information about both individual identities and personal familiarity, while clarifying and extending the organization of the core system for face perception.

Concurrent development of facial identity and expression discrimination.

Facial identity and facial expression processing both appear to follow a protracted developmental trajectory, yet these trajectories have been studied independently and have not been directly compared. Here we investigated whether these processes develop at the same or different rates using matched identity and expression discrimination tasks. The Identity task begins with a target face that is a morph between two identities (Identity A/Identity B). After a brief delay, the target face is replaced by two choice faces: 100% Identity A and 100% Identity B. Children 5-12-years-old were asked to pick the choice face that is most similar to the target identity. The Expression task is matched in format and difficulty to the Identity task, except the targets are morphs between two expressions (Angry/Happy, or Disgust/Surprise). The same children were asked to pick the choice face with the expression that is most similar to the target expression. There were significant effects of age, with performance improving (becoming more accurate and faster) on both tasks with increasing age. Accuracy and reaction times were not significantly different across tasks and there was no significant Age x Task interaction. Thus, facial identity and facial expression discrimination appear to develop at a similar rate, with comparable improvement on both tasks from age five to twelve. Because our tasks are so closely matched in format and difficulty, they may prove useful for testing face identity and face expression processing in special populations, such as autism or prosopagnosia, where one of these abilities might be impaired.

Familiarity facilitates feature-based face processing.

Recognition of personally familiar faces is remarkably efficient, effortless and robust. We asked if feature-based face processing facilitates detection of familiar faces by testing the effect of face inversion on a visual search task for familiar and unfamiliar faces. Because face inversion disrupts configural and holistic face processing, we hypothesized that inversion would diminish the familiarity advantage to the extent that it is mediated by such processing. Subjects detected personally familiar and stranger target faces in arrays of two, four, or six face images. Subjects showed significant facilitation of personally familiar face detection for both upright and inverted faces. The effect of familiarity on target absent trials, which involved only rejection of unfamiliar face distractors, suggests that familiarity facilitates rejection of unfamiliar distractors as well as detection of familiar targets. The preserved familiarity effect for inverted faces suggests that facilitation of face detection afforded by familiarity reflects mostly feature-based processes.