Face Memory Research Articles

Long-term memory for faces in dysmorphic concern and self-reported body dysmorphic disorder

Dysmorphic concerns are a core symptom of body dysmorphic disorder (BDD), a mental disorder characterized by a preoccupation with perceived appearance-related flaws. Different cognitive and perceptual biases are associated with dysmorphic concern or BDD, including object-related short-term memory deficits and superior short-term memory performance for facial features. A face-related long-term memory bias might explain the feature of perceiving flaws and why individuals with BDD experience adverse outcomes of cosmetic surgery. The present study aimed at investigating long-term memory (LTM) performance for different facial stimuli. 62 participants completed an online experiment (30 with low and 32 with high dysmorphic concern; 39 without and 23 with self-reported BDD). The experiment consisted of two consecutive old/new recognition tasks, testing memory performance of whole faces and facial features. LTM performance was higher for whole faces than for facial features across groups. Neither significant differences between the low and high dysmorphic concern groups nor an interaction of group x stimulus type were found. Post hoc exploratory analyses revealed lower long-term memory performance in individuals with versus without self-reported BDD. The results indicate that memory performance might be only affected in BDD rather than non-clinical or non–BDD-specific dysmorphic concerns.

Exploring the Efficacy of a Remote Strategy-Based Intervention for People With Multiple Sclerosis With Everyday Memory Impairments: A Pilot Study.

This pilot study evaluates a remote strategy-based intervention for individuals with multiple sclerosis who experience everyday memory impairments. The intervention can potentially inform cognitive rehabilitation for this population. To investigate the feasibility and efficacy of an intervention (TELE-Self-GEN) to determine whether it can alleviate everyday memory impairments of individuals with multiple sclerosis. Pretest-posttest. Community. Ten adults with multiple sclerosis. Six synchronous treatment sessions were delivered online via Zoom. The treatment protocol embedded a memory strategy (self-generated learning) within a metacognitive framework, including self-awareness and self-management strategies. The treatment emphasizes when and how self-generation should be used. Measurements assessed feasibility and participants' satisfaction with the intervention and its delivery method, as well as memory, everyday memory, and functional performance. Participants expressed high satisfaction with the virtual treatment, highlighting its convenience as a key factor. Treatment resulted in improvements in memory performance, perceived memory ability in daily life, and functional performance. Results provide initial proof of concept in the utilization of a remotely delivered, strategy-based treatment approach to improve memory performance and functional abilities. The pilot data support a larger randomized clinical trial of the TELE-self-GEN. Plain-Language Summary: The results of this pilot study highlight the promising potential of TELE-self-GEN for people with multiple sclerosis (MS), who face memory challenges every day. This remotely delivered, strategy-based occupational therapy treatment approach, TELE-self-GEN, has the potential to significantly improve functional memory. The study participants reported improvements in their memory performance, perceived memory ability in daily life, and functional performance. These encouraging results serve as a foundation for more extensive clinical trials using TELE-self-GEN for people with MS.

B-move: faster bidirectional character extensions in a run-length compressed index.

Due to the increasing availability of high-quality genome sequences, pan-genomes are gradually replacing single consensus reference genomes in many bioinformatics pipelines to better capture genetic diversity. Traditional bioinformatics tools using the FM-index face memory limitations with such large genome collections. Recent advancements in run-length compressed indices like Gagie et al.'s r-index and Nishimoto and Tabei's move structure, alleviate memory constraints but focus primarily on backward search for MEM-finding. Arakawa et al.'s br-index initiates complete approximate pattern matching using bidirectional search in run-length compressed space, but with significant computational overhead due to complex memory access patterns. We introduce b-move, a novel bidirectional extension of the move structure, enabling fast, cache-efficient bidirectional character extensions in run-length compressed space. It achieves bidirectional character extensions up to 8 times faster than the br-index, closing the performance gap with FM-index-based alternatives, while maintaining the br-index's favorable memory characteristics. For example, all available complete E. coli genomes on NCBI's RefSeq collection can be compiled into a b-move index that fits into the RAM of a typical laptop. Thus, b-move proves practical and scalable for pan-genome indexing and querying. We provide a C++ implementation of b-move, supporting efficient lossless approximate pattern matching including locate functionality, available at https://github.com/biointec/b-move under the AGPL-3.0 license.

Robust memory of face moral values is encoded in the human caudate tail: a simultaneous EEG-fMRI study

Moral judgements about people based on their actions is a key component that guides social decision making. It is currently unknown how positive or negative moral judgments associated with a person’s face are processed and stored in the brain for a long time. Here, we investigate the long-term memory of moral values associated with human faces using simultaneous EEG-fMRI data acquisition. Results show that only a few exposures to morally charged stories of people are enough to form long-term memories a day later for a relatively large number of new faces. Event related potentials (ERPs) showed a significant differentiation of remembered good vs bad faces over centerofrontal electrode sites (value ERP). EEG-informed fMRI analysis revealed a subcortical cluster centered on the left caudate tail (CDt) as a correlate of the face value ERP. Importantly neither this analysis nor a conventional whole-brain analysis revealed any significant coding of face values in cortical areas, in particular the fusiform face area (FFA). Conversely an fMRI-informed EEG source localization using accurate subject-specific EEG head models also revealed activation in the left caudate tail. Nevertheless, the detected caudate tail region was found to be functionally connected to the FFA, suggesting FFA to be the source of face-specific information to CDt. A further psycho-physiological interaction analysis also revealed task-dependent coupling between CDt and dorsomedial prefrontal cortex (dmPFC), a region previously identified as retaining emotional working memories. These results identify CDt as a main site for encoding the long-term value memories of faces in humans suggesting that moral value of faces activates the same subcortical basal ganglia circuitry involved in processing reward value memory for objects in primates.

Facial expressions affect the memory of facial colors.

Facial color influences the perception of facial expressions, and emotional expressions bias how facial color is remembered. However, it remains unclear whether facial expressions affect daily facial color memory. The memory color effect demonstrates that knowledge about typical colors affects the perception of the actual color of given objects. To investigate the effect of facial color memory, we examined whether the memory color effect for faces varies depending on facial expression. We calculated the subjective achromatic point of the facial expression image stimulus and compared the degree to which it was shifted from the actual achromatic point between facial expression conditions. We hypothesized that if the memory of facial color is influenced by the facial expression color (e.g., anger is a warm color, fear is a cold color), then the subjective achromatic point would vary with facial expression. In Experiment 1, we recruited 13 participants who adjusted the color of facial expression stimuli (anger, neutral, and fear) and a banana stimulus to be achromatic. No significant differences in the subjective achromatic point between facial expressions were observed. Subsequently, we conducted Experiment 2 with 23 participants because Experiment 1 did not account for the sensitivity to color changes on the face; humans perceive greater color differences in faces than in non-faces. Participants selected which facial color they believed the expression stimulus appeared to be, choosing one of two options provided to them. The results indicated that the subjective achromatic points of anger and fear faces significantly shifted toward the opposite color direction compared with neutral faces in the brief presentation condition. This research suggests that the memory color of faces differs depending on facial expressions and supports the idea that the perception of emotional expressions can bias facial color memory.

Temporal multiplexing of perception and memory codes in IT cortex

A central assumption of neuroscience is that long-term memories are represented by the same brain areas that encode sensory stimuli1. Neurons in inferotemporal (IT) cortex represent the sensory percept of visual objects using a distributed axis code2–4. Whether and how the same IT neural population represents the long-term memory of visual objects remains unclear. Here we examined how familiar faces are encoded in the IT anterior medial face patch (AM), perirhinal face patch (PR) and temporal pole face patch (TP). In AM and PR we observed that the encoding axis for familiar faces is rotated relative to that for unfamiliar faces at long latency; in TP this memory-related rotation was much weaker. Contrary to previous claims, the relative response magnitude to familiar versus unfamiliar faces was not a stable indicator of familiarity in any patch5–11. The mechanism underlying the memory-related axis change is likely intrinsic to IT cortex, because inactivation of PR did not affect axis change dynamics in AM. Overall, our results suggest that memories of familiar faces are represented in AM and perirhinal cortex by a distinct long-latency code, explaining how the same cell population can encode both the percept and memory of faces.

The effect of misophonia on cognitive and social judgments.

Misophonia, a heightened aversion to certain sounds, turns common cognitive and social exercises (e.g., paying attention during a lecture near a pen-clicking classmate, coexisting at the dinner table with a food-chomping relative) into challenging endeavors. How does exposure to triggering sounds impact cognitive and social judgments? We investigated this question in a sample of 65 participants (26 misophonia, 39 control) from the general population. In Phase 1, participants saw faces paired with auditory stimuli while completing a gender judgment task, then reported sound discomfort and identification. In Phase 2, participants saw these same faces with novel ones and reported face likeability and memory. For both oral and non-oral triggers, misophonic participants gave higher discomfort ratings than controls did-especially when identification was correct-and performed slower on the gender judgment. Misophonic participants rated lower likeability than controls did for faces they remembered with high discomfort sounds, and face memory was worse overall for faces originally paired with high discomfort sounds. Altogether, these results suggest that misophonic individuals show impairments on social and cognitive judgments if they must endure discomforting sounds. This experiment helps us better understand the day-to-day impact of misophonia and encourages usage of individualized triggers in future studies.

Pimo: memory-efficient privacy protection in video streaming and analytics

Video streaming from cameras to backend cloud or edge servers for neural-based analytics has gained significant popularity. However, the transmission of data from cameras to a backend raises substantial privacy concerns, particularly regarding sensitive information like facial data. To offer privacy protection, visual processing techniques, such as Generative Adversarial Networks (GANs), have been employed on cameras to blur and safeguard such data intelligently. However, these techniques frequently face memory challenges, particularly when dealing with high-resolution videos. In this paper, we propose PIMO, a memory-efficient visual privacy protection scheme designed to effectively blur video content leveraging adaptive slicing of frames and resolution degradation. Our extensive experimental evaluations validate that PIMO’s adaptive mechanism proficiently navigates fluctuating memory constraints. Furthermore, utilizing a content-based blur scheme, our approach can maintain an impressive mean precision of 95.2%, as compared to the original, non-blurred images.

Perception and Memory Reinstatement Engage Overlapping Face-Selective Regions within Human Ventral Temporal Cortex.

Humans have the remarkable ability to vividly retrieve sensory details of past events. According to the theory of sensory reinstatement, during remembering, brain regions specialized for processing specific sensory stimuli are reactivated to support content-specific retrieval. Recently, several studies have emphasized transformations in the spatial organization of these reinstated activity patterns. Specifically, studies of scene stimuli suggest a clear anterior shift in the location of retrieval activations compared with the activity observed during perception. However, it is not clear that such transformations occur universally, with inconsistent evidence for other important stimulus categories, particularly faces. One challenge in addressing this question is the careful delineation of face-selective cortices, which are interdigitated with other selective regions, in configurations that spatially differ across individuals. Therefore, we conducted a multisession neuroimaging study to first carefully map individual participants' (nine males and seven females) face-selective regions within ventral temporal cortex (VTC), followed by a second session to examine the activity patterns within these regions during face memory encoding and retrieval. While face-selective regions were expectedly engaged during face perception at encoding, memory retrieval engagement exhibited a more selective and constricted reinstatement pattern within these regions, but did not show any consistent direction of spatial transformation (e.g., anteriorization). We also report on unique human intracranial recordings from VTC under the same experimental conditions. These findings highlight the importance of considering the complex configuration of category-selective cortex in elucidating principles shaping the neural transformations that occur from perception to memory.

Пространственно-ориентационные параметры узнавания изображений лиц как предикторы симптомов печеночной недостаточности у подростков

Functional assessment of higher mental functions in case of intoxication or during treatment will make it possible to identify predictors of the symptoms of hepatic encephalopathy associated with portal hypertension. The study was aimed to determine the diagnostic predictors of the emergence of the earliest symptoms of hepatic encephalopathy in adolescents with portal hypertension. The study involved 60 adolescents aged 13–17 years: 28 males, 32 females. The experimental group included 30 adolescents with the diagnosis К76.6 Portal hypertension, unspecified form. The control group included 30 adolescents with normal somatic status, who had no mental disorders, traumatic brain injuries or severe infectious diseases of the brain (based on the records of the annual check-up). The studied groups were matched by sex, age, and social status. The proprietary method, Tobii EyeX hardware and software system (GazeControl software), and Cambridge Face Memory Test for Children (CFMT-C) were used. It has been found that recognition of single face images, multiple face images, and multiple face images camouflaged with noise by adolescents with portal hypertension is associated with the greater efforts (manifested in the increased number and duration of gaze fixations), than recognition of the above by adolescents with normal somatic status. The accuracy of recognition of single face images, multiple face images, and multiple face images camouflaged with noise shown by adolescents with portal hypertension experiencing the toxic effects associated with manifestations of hepatic encephalopathy is significantly reduced compared to that shown by adolescents with normal somatic status. The results obtained can be considered as the diagnostic predictors allowing one to trace the changes in the hepatic encephalopathy severity at various stages of treatment (including after surgical intervention).

Relationship between face recognition ability and anxiety tendencies in healthy young individuals: A prosopagnosia index and state-trait anxiety inventory study

Developmental prosopagnosia (DP) is a condition that indicates the inability to recognize individuals by their faces from birth, without any history of brain damage. The assessment of face recognition ability and diagnosis of DP involve the use of face tests such as the Cambridge Face Memory Test (CFMT) and the Cambridge Face Perception Test, along with self-reported measures like the 20-Item Prosopagnosia Index (PI20). Face recognition accuracy is affected by anxiety. However, previous studies on the relationship between face recognition ability and anxiety have not used the PI20 measure. This study aimed to investigate the relationship between self-reported measures of face recognition ability and anxiety tendencies among healthy young individuals for DP diagnosis and its implications. We used a face recognition test, involving the PI20, CFMT, Visual Perception Test for Agnosia–Famous Face Test (VPTA–FFT), and State-Trait Anxiety Inventory (STAI). We assessed the performance of 116 Japanese young adults (75 females, median age of 20.7 years, with a standard deviation of 1.2). Subsequently, we conducted a statistical analysis to examine the relationship between the outcomes of the face recognition tests and STAI scores using Pearson correlation analysis and single correlation coefficients. The results showed a positive correlation between state anxiety and PI20 (r = 0.308, p = 0.007), and a weak positive correlation was also observed between trait anxiety and PI20 (r = 0.268, p = 0.04). In contrast, there was no correlation between CFMT and VPTA–FFT with respect to STAI. The results of the hierarchical multiple regression analysis also suggested that the correlation between the performance on the PI20 (self-report) and objective measures of face recognition performance (the CFMT and the VPTA–FFT) are driven by differences in anxiety. This study is the first to explore the relationship between face recognition abilities and anxiety using the PI20 self-report measure. There are implications for future research on the diagnosis of DP and the relationship between anxiety and face recognition.

Both identity and non-identity face perception tasks predict developmental prosopagnosia and face recognition ability

Developmental prosopagnosia (DP) is characterised by deficits in face identification. However, there is debate about whether these deficits are primarily perceptual, and whether they extend to other face processing tasks (e.g., identifying emotion, age, and gender; detecting faces in scenes). In this study, 30 participants with DP and 75 controls completed a battery of eight tasks assessing four domains of face perception (identity; emotion; age and gender; face detection). The DP group performed worse than the control group on both identity perception tasks, and one task from each other domain. Both identity perception tests uniquely predicted DP/control group membership, and performance on two measures of face memory. These findings suggest that deficits in DP may arise from issues with face perception. Some non-identity tasks also predicted DP/control group membership and face memory, even when face identity perception was accounted for. Gender perception and speed of face detection consistently predicted unique variance in group membership and face memory; several other tasks were only associated with some measures of face recognition ability. These findings indicate that face perception deficits in DP may extend beyond identity perception. However, the associations between tasks may also reflect subtle aspects of task demands or stimuli.

The Neural Mechanism Underlying Differentiated In-Group Versus Out-Group Face Recognition and Memory, Identification, Empathy and Pro-social Behavior: Evidence from fMRI and ERP Studies.

In the context of perceiving individuals within and outside of social groups, there are distinct cognitive processes and mechanisms in the brain. Extensive research in recent years has delved into the neural mechanisms that underlie differences in how we perceive individuals from different social groups. To gain a deeper understanding of these neural mechanisms, we present a comprehensive review from the perspectives of facial recognition and memory, intergroup identification, empathy, and pro-social behavior. Specifically, we focus on studies that utilize functional magnetic resonance imaging (fMRI) and event-related potential (ERP) techniques to explore the relationship between brain regions and behavior. Findings from fMRI studies reveal that the brain regions associated with intergroup differentiation in perception and behavior do not operate independently but instead exhibit dynamic interactions. Similarly, ERP studies indicate that the amplitude of neural responses shows various combinations in relation to perception and behavior.

Glucose Metabolism of Hippocampal Subfields in Medial Temporal Lobe Epilepsy.

Reduced glucose metabolism in the hippocampus is commonly observed in cases of medial temporal lobe epilepsy (MTLE) with hippocampal sclerosis (HS). Glucose metabolism among the various hippocampal subfields has not been thoroughly investigated. This study examined 29 patients (18 females; 15-58 years) diagnosed with HS who underwent surgery for drug-resistant epilepsy. FreeSurfer 7.1.1 was used in the processing of MRI data and 18 F-FDG PET scans to derive volumetric data and the FDG SUVr in the whole hippocampus and hippocampal subfields, including the CA1, CA2-4, granule cell and molecular layer of the dentate gyrus (GC-ML-DG), and subiculum. Asymmetries in the volume and SUVr between the 2 sides from the subfields of the hippocampus were defined in terms of an asymmetry index. Comparisons of the asymmetry index among these regions were performed. The correlations between asymmetry index values and postoperative outcomes and presurgical neuropsychological test results were also evaluated. The CA1, CA2-4, subiculum, GC-ML-DG, and whole hippocampus presented reductions in volume and hypometabolism ipsilateral to MTLE. Asymmetries in volume and SUVr were significantly less pronounced in the CA1 and subiculum than in the CA2-4 or GC-ML-DG. Postoperative seizure outcomes were not correlated with the asymmetry index for volume or SUVr in any hippocampal subfield. In cases of left MTLE, scores of immediate logical memory and delayed logical memory were positively correlated with the asymmetry index for SUVr in the following subfields: CA1 ( R = 0.829, P = 0.021; R = 0.770, P = 0.043), CA2-4 ( R = 0.825, P = 0.022; R = 0.894, P = 0.007), subiculum ( R = 0.882, P = 0.009; R = 0.853, P = 0.015), GC-ML-DG ( R = 0.850, P = 0.015; R = 0.796, P = 0.032), and whole hippocampus ( R = 0.841, P = 0.018; R = 0.822, P = 0.023). In cases of right MTLE, the scores for delayed face memory were positively correlated with the asymmetry index for SUVr in the subiculum ( R = 0.935, P = 0.006). In cases of HS, changes in glucose metabolism levels varied among the hippocampal subfields. Asymmetries in glucose metabolism among the CA-1, CA2-4, subiculum, and GC-ML-DG subregions were correlated with scores for verbal memory among patients with left MTLE. Asymmetric glucose metabolism in the subiculum was also correlated with visual memory scores among patients with right MTLE.

Is “memory-for-when” universal? Group and individual variability in temporal position memory for words, faces, and classrooms

ABSTRACT Many contemporary theories of memory assume that everyone automatically stores temporal contextual information about all types of encountered information, yet most studies on this topic have used words and ignored individual differences. Five experiments accumulated evidence that explicit storage of temporal context information does not appear to occur automatically for all people and types of memoranda. We collected judgments of temporal position (memory-for-when) for words (Experiments 1 & 3), faces (Experiments 2A, 3, 4, and 5), and classrooms (Experiments 2B & 3). At the group level, for each of these memoranda memory-for-when was sensitive to the original input position and showed a temporal primacy effect reflecting better memory for position for items near the beginning of the list, indicating some automatic storage of temporal context information. However, memory-for-when was significantly better for words than classrooms, with faces in the middle. Moreover, individuals varied dramatically in their ability to indicate memory-for-when, especially for classrooms where many people performed at or near chance. Taken together, the data suggest that explicit memory-for-when may be dissociable from the more implicit use of temporal contextual information that is theorised to occur during free recall.

Comparison with others influences encoding and recognition of their faces: Behavioural and ERP evidence

In daily life, faces are often memorized within contexts involving interpersonal interactions. However, little is known about whether interpersonal interaction-related contexts influence face memory. The present study aimed to understand this question by investigating how social comparison-related context affects face encoding and recognition. To address this issue, 40 participants were informed that they and another player each played a monetary game and were then presented with both of their outcomes (either monetary gain or loss). Subsequently, participants were shown the face of the player whom they were just paired with. After all the faces had been encoded, participants were asked to perform a sudden old/new recognition task involving these faces. The results showed that, during the encoding phase, another player's monetary gain, compared to loss, resulted in more negative responses in the N170 and early posterior negativity (EPN)/N250 to relevant players’ faces when participants encountered monetary loss and a smaller late positive potential (LPP) response irrespective of self-related outcomes. In the subsequent recognition phase, preceding another player's monetary gain as compared to loss led to better recognition performance and stronger EPN/N250 and LPP responses to the faces of relevant players when participants had lost some amount of money. These findings suggest that the social comparison-related context, particularly self-disadvantageous outcomes in the context, influences the memory of comparators’ faces.

The effect of reward expectation on working memory of emotional faces under different levels of cognitive load: an ERP study.

Using event-related potentials (ERPs), this study examined the impact of reward expectations on working memory of emotional faces under different levels of cognitive load in a task combining the N-back paradigm and the reward expectation paradigm. The experiment involved presenting high- or low-reward cues followed by an N-back task for emotional faces with different loads. The accuracy results showed that under a high task load, both reward and emotion effects were significantly observed. However, these effects disappeared under a low task load. Analysis of the ERP data revealed that the early P2 and VPP components exhibited greater responses to fearful faces than to neutral faces. In the later stages, the P3 and LPP components showed greater reactions to high rewards than to low rewards. Additionally, the P2 component was found to be modulated by task load in relation to rewards, the EPN component demonstrated task load modulation with respect to emotions, and the N170 component showed an interaction effect between rewards and emotions. These findings imply that load regulates the reward effect and the emotional superiority effect in the process of working memory for emotional faces. In the cognitive processing of working memory, motivation and emotion jointly influence processing. Emotional factors have a greater impact in the early stage of processing, while motivation factors have a greater impact in the late stage of processing.

Trial-level fluctuations in pupil dilation at encoding reflect strength of relational binding.

Eye-tracking methodologies have revealed that eye movements and pupil dilations are influenced by our previous experiences. Dynamic fluctuations in pupil size during learning reflect in part the formation of memories for learned information, while viewing behavior during memory testing is influenced by memory retrieval and drawn to previously learned associations. However, no study to date has linked fluctuations in pupil dilation at encoding to the magnitude of viewing behavior at test. The current investigation involved monitoring eye movements both in single item recognition and relational recognition tasks. In the item task, all faces were presented with the same background scene and memory for faces was subsequently tested, whereas in the relational task each face was presented with its own unique background scene and memory for the face-scene association was subsequently tested. Pupil size changes during encoding predicted the magnitude of preferential viewing during test, as well as future recognition accuracy. These effects emerged only in the relational task, but not in the item task, and were replicated in an additional experiment in which stimulus luminance was more tightly controlled. A follow-up experiment and additional analyses ruled out differences in orienting instructions or number of fixations to the encoding display as explanations of the observed effects. The results shed light on the links between pupil dilation, memory encoding, and eye movement patterns during recognition and suggest that trial-level fluctuations in pupil dilation during encoding reflect relational binding of items to their context rather than general memory formation or strength. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Generalized Anxiety Disorder and Depression Associated with Developmental Prosopagnosia: A Case Report

Developmental prosopagnosia is a disorder of facial recognition that begins during early childhood in the absence of acquired central nervous system disease. We report the case of a 15-year-old female with developmental prosopagnosia as measured by the 20-item Prosopagnosia Index and Cambridge Face Memory Test who ultimately developed generalized anxiety disorder and depression despite relatively normal social and psychological function during early childhood. In elementary school, the case patient adapted by learning alternative ways to identify others, such as by clothing and hairstyle, but this became more difficult in junior high school due to the requirement for school uniforms and regulations on hairstyle. This difficulty in turn led to interpersonal problems that ultimately resulted in symptoms of generalized anxiety disorder and depression, such as headache and sleep dysfunction. People with developmental prosopagnosia are generally prone to having depressed and anxious feelings. However, to the best of our knowledge, this is the first case report of anxiety disorder or depression related to developmental prosopagnosia. This comorbidity may be relatively common, especially in ethnically homogeneous countries with strict school regulations on personal appearance such as Japan.

Neural mechanisms of face familiarity and learning in the human amygdala and hippocampus

Recognizing familiar faces and learning new faces play an important role in social cognition. However, the underlying neural computational mechanisms remain unclear. Here, we record from single neurons in the human amygdala and hippocampus and find a greater neuronal representational distance between pairs of familiar faces than unfamiliar faces, suggesting that neural representations for familiar faces are more distinct. Representational distance increases with exposures to the same identity, suggesting that neural face representations are sharpened with learning and familiarization. Furthermore, representational distance is positively correlated with visual dissimilarity between faces, and exposure to visually similar faces increases representational distance, thus sharpening neural representations. Finally, we construct a computational model that demonstrates an increase in the representational distance of artificial units with training. Together, our results suggest that the neuronal population geometry, quantified by the representational distance, encodes face familiarity, similarity, and learning, forming the basis of face recognition and memory.