Ph. D. Project
Title:
Cellular mechanisms of facial identity coding in humans
Dates:
2020/10/01 - 2023/09/30
Supervisor(s):
Description:
The objective of this PhD proposal is to characterize and understand how face identity is coded at the level of small populations of neurons distributed in the ventral occipito-temporal cortex, in interaction with medial temporal lobe structures, in the human brain.
The face is arguably the most widely studied stimulus in cognitive neuroscience. Why? Because faces are highly frequent stimuli in the environment, especially in modern societies and in the age of (social) media. In humans, faces provide a wealth of information, e.g., emotion through expression, direction of attention from head and gaze position, identity, age, sex, ethnical origin, attractiveness, and personality traits. Further, it is a complex visual stimulus, containing multiple nameable parts (eyes, mouth, ...), being therefore well suited to study perception in a modality, vision, dominant in the human species. Moreover, faces form a rich visual category, including a wide range of physically variable face exemplars but excluding physically similar non-face objects. This is one of the reasons why there is a large amount of research in artificial intelligence for automatic face recognition systems that, nevertheless, cannot approach human performance at face recognition. Various neurological (Alzheimer's disease, semantic dementia, ...) and psychiatric (autism, ...) disorders have also been linked to deficiencies at interpreting signals from the face, making it a well-suited stimulus to better understand these disorders.
Lesion analysis, neuroimaging and intracranial recordings have shown that a large expanse of the human cerebral cortex is involved in face recognition, in part specifically, and with a right hemispheric lateralization. Understanding face recognition is therefore of primary importance to understand the general organization of cognitive functions in the human brain, including their hemispheric lateralization and cortical specialization. The most challenging function to understand is the recognition of facial identity: young neurotypical adults living in large societies are able to recognize up to 5000 identities from their face (Jenkins et al., 2018), which is highly challenging because individual faces can be very similar, and individual faces change substantially across viewing conditions. Unfortunately, how does the brain encode so many distinctive face patterns and is able to generalize its response to variable view of a single identity is a scientific mystery, preventing also to understand how and why the human brain sometimes forgets or fails to recognize face identity. To understand these mechanisms, the neuroscientific community has largely relied on neurophysiological studies performed in the macaque monkey (e.g., Chang & Tsao, 2017).
However, unlike humans, macaques are not very efficient at face identity recognition and do not have the humanlike neuro-anatomical structures to carry this function, making this animal model largely inadequate (Rossion & Taubert, 2019). Clarifying how face identity is coded at the level small populations of neurons requires direct recording in the human ventral occipito-temporal cortex (VOTC). In humans, recordings the activity of neurons is seriously limited, being performed only in a few clinical centers in patients with refractory epilepsy implanted with intracerebral electrodes. Following years of development and thanks to equipment supported initially by the Lorraine Region, this approach is now available and functional at the epilepsy unit of the CHRU-Nancy, in close collaboration with researchers in neuroscience and signal processing of the CRAN BIOSIS. The PhD candidate will record multi-unit and single-unit activity in responses to faces in human brain structures, with a focus on fMRI-defined face-selective regions of the ventral occipito-temporal cortex (VOTC). Recordings will also be performed in the medial temporal lobe (MTL; hippocampus, amygdala), where neurons firing to multimodal constructs have been found (e.g., "concept neurons", e.g., "Jennifer Aniston neuron", Quiroga, 2012). He/she will use and develop approaches based on the principle of frequency-tagging (Norcia et al., 2015) to isolate neuronal responses to faces as compared to other stimuli, the discrimination between unfamiliar individual faces, or between familiar and unfamiliar face identities. The PhD candidate will analyse and characterize multi-unit and single-unit neuronal data in the frequency-domain and in the time-domain. Single unit activity will be isolated using locally and internationally developed spike sorting algorithm (in collaboration with R. Quiroga, Leicester University, UK). The comparison of the nature and spatio-temporal characteristics of the responses recorded in several structures of the VOTC, and between the MTL and the VOTC, will help making decisive progress in our understanding of how individual faces are (en)coded in the human brain. The effect of focal (e.g., Jonas et al., 2018) and global (tDCS) electrical stimulation will be used to test how the coding of face identity in local regions can be disrupted, but also evaluate functional interactions between the VOTC and MTL during facial identity discrimination and memory encoding.
Norcia, A.M., Appelbaum, G., Ales, J., Cottereau, B., Rossion, B. (2015). The Steady-State Visual Evoked Potential in Vision Research: a Review. Journal of Vision, 15(6):4, 1-46.
Chang, L., & Tsao, D. Y. (2017). The code for facial identity in the primate brain. Cell, 169,
1013-1028.
Jenkins, R., Dowsett A.J., Burton, A.M. (2018). How many faces do people know? Proc Biol Sci., 285 (1888).
Rossion, B. & Taubert, J. (2019). What can we learn about human individual face recognition from experimental studies in monkeys? Vision Research, 157, 142-158.
Jonas, J., Brissart, H., Hossu, G., Colnat-Coulbois, S., Vignal, J.-P., Rossion, B., Maillard, L. (2018). A face identity hallucination (palinopsia) generated by intracerebral stimulation of the face-selective right lateral fusiform cortex. Cortex, 99, 296-310.
Quiroga, R.Q. (2012) Concept cells: the building blocks of declarative memory functions. Nat Rev Neurosci., 13, 587-97.
The face is arguably the most widely studied stimulus in cognitive neuroscience. Why? Because faces are highly frequent stimuli in the environment, especially in modern societies and in the age of (social) media. In humans, faces provide a wealth of information, e.g., emotion through expression, direction of attention from head and gaze position, identity, age, sex, ethnical origin, attractiveness, and personality traits. Further, it is a complex visual stimulus, containing multiple nameable parts (eyes, mouth, ...), being therefore well suited to study perception in a modality, vision, dominant in the human species. Moreover, faces form a rich visual category, including a wide range of physically variable face exemplars but excluding physically similar non-face objects. This is one of the reasons why there is a large amount of research in artificial intelligence for automatic face recognition systems that, nevertheless, cannot approach human performance at face recognition. Various neurological (Alzheimer's disease, semantic dementia, ...) and psychiatric (autism, ...) disorders have also been linked to deficiencies at interpreting signals from the face, making it a well-suited stimulus to better understand these disorders.
Lesion analysis, neuroimaging and intracranial recordings have shown that a large expanse of the human cerebral cortex is involved in face recognition, in part specifically, and with a right hemispheric lateralization. Understanding face recognition is therefore of primary importance to understand the general organization of cognitive functions in the human brain, including their hemispheric lateralization and cortical specialization. The most challenging function to understand is the recognition of facial identity: young neurotypical adults living in large societies are able to recognize up to 5000 identities from their face (Jenkins et al., 2018), which is highly challenging because individual faces can be very similar, and individual faces change substantially across viewing conditions. Unfortunately, how does the brain encode so many distinctive face patterns and is able to generalize its response to variable view of a single identity is a scientific mystery, preventing also to understand how and why the human brain sometimes forgets or fails to recognize face identity. To understand these mechanisms, the neuroscientific community has largely relied on neurophysiological studies performed in the macaque monkey (e.g., Chang & Tsao, 2017).
However, unlike humans, macaques are not very efficient at face identity recognition and do not have the humanlike neuro-anatomical structures to carry this function, making this animal model largely inadequate (Rossion & Taubert, 2019). Clarifying how face identity is coded at the level small populations of neurons requires direct recording in the human ventral occipito-temporal cortex (VOTC). In humans, recordings the activity of neurons is seriously limited, being performed only in a few clinical centers in patients with refractory epilepsy implanted with intracerebral electrodes. Following years of development and thanks to equipment supported initially by the Lorraine Region, this approach is now available and functional at the epilepsy unit of the CHRU-Nancy, in close collaboration with researchers in neuroscience and signal processing of the CRAN BIOSIS. The PhD candidate will record multi-unit and single-unit activity in responses to faces in human brain structures, with a focus on fMRI-defined face-selective regions of the ventral occipito-temporal cortex (VOTC). Recordings will also be performed in the medial temporal lobe (MTL; hippocampus, amygdala), where neurons firing to multimodal constructs have been found (e.g., "concept neurons", e.g., "Jennifer Aniston neuron", Quiroga, 2012). He/she will use and develop approaches based on the principle of frequency-tagging (Norcia et al., 2015) to isolate neuronal responses to faces as compared to other stimuli, the discrimination between unfamiliar individual faces, or between familiar and unfamiliar face identities. The PhD candidate will analyse and characterize multi-unit and single-unit neuronal data in the frequency-domain and in the time-domain. Single unit activity will be isolated using locally and internationally developed spike sorting algorithm (in collaboration with R. Quiroga, Leicester University, UK). The comparison of the nature and spatio-temporal characteristics of the responses recorded in several structures of the VOTC, and between the MTL and the VOTC, will help making decisive progress in our understanding of how individual faces are (en)coded in the human brain. The effect of focal (e.g., Jonas et al., 2018) and global (tDCS) electrical stimulation will be used to test how the coding of face identity in local regions can be disrupted, but also evaluate functional interactions between the VOTC and MTL during facial identity discrimination and memory encoding.
Norcia, A.M., Appelbaum, G., Ales, J., Cottereau, B., Rossion, B. (2015). The Steady-State Visual Evoked Potential in Vision Research: a Review. Journal of Vision, 15(6):4, 1-46.
Chang, L., & Tsao, D. Y. (2017). The code for facial identity in the primate brain. Cell, 169,
1013-1028.
Jenkins, R., Dowsett A.J., Burton, A.M. (2018). How many faces do people know? Proc Biol Sci., 285 (1888).
Rossion, B. & Taubert, J. (2019). What can we learn about human individual face recognition from experimental studies in monkeys? Vision Research, 157, 142-158.
Jonas, J., Brissart, H., Hossu, G., Colnat-Coulbois, S., Vignal, J.-P., Rossion, B., Maillard, L. (2018). A face identity hallucination (palinopsia) generated by intracerebral stimulation of the face-selective right lateral fusiform cortex. Cortex, 99, 296-310.
Quiroga, R.Q. (2012) Concept cells: the building blocks of declarative memory functions. Nat Rev Neurosci., 13, 587-97.
Keywords:
face identity, single unit recording, neurons, cortex, electrical stimulation
Conditions:
Duration: 3 years
Employer: CRAN UMR7039
Localization: CRAN UMR7039, Cognitive and Systems Neuroscience division, Pavillon Krug, Hopital Central, CHRU-Nancy
Salary: 92 000¬ for 3 years
Master degree (or equivalent) in Neuroscience, Cognitive Science, Psychology, MedecineMedicine, or Biomedical Engineering.
Employer: CRAN UMR7039
Localization: CRAN UMR7039, Cognitive and Systems Neuroscience division, Pavillon Krug, Hopital Central, CHRU-Nancy
Salary: 92 000¬ for 3 years
Master degree (or equivalent) in Neuroscience, Cognitive Science, Psychology, MedecineMedicine, or Biomedical Engineering.
Department(s):
| Biology, Signals and Systems in Cancer and Neuroscience |
Funds:
Doctoral fellowship UL, co-supported by Région Grand Est
Publications:
doi: 10.1111/nyas.13596; doi: 10.1073/pnas.1522033113; doi: 10.1016/j.neuroimage.2014.06.017 + CRAN - Publications
