Research ReportDifferent and common brain signals of altered neurocognitive mechanisms for unfamiliar face processing in acquired and developmental prosopagnosia
Introduction
One of the most admirable human skills in social communication is the ability to visually distinguish and recognise hundreds and even thousands of faces that are met across lifespan. This ability requires the recruitment of subtle perceptual operations, including the integration of facial features into global face configurations, in order to (potentially) create adequate face representations in memory (Bruce, 1988; Maurer, Le Grand, & Mondloch, 2002; Schwaninger, Lobmaier & Collishaw, 2002; Sergent, 1986; Tanaka & Farah, 1993). How this integration is carried out is so far an insufficiently investigated topic in cognitive neuroscience. From a neuropsychological perspective, some studies have found that prosopagnosic people, those individuals with atypical difficulties to correctly perceive or recognise faces (Damasio, Damasio, & Vanhoesen, 1982; De Renzi, 1986; Ellis & Florence, 1990; Hécaen & Angelergues, 1962), are impaired in the perception and recognition of both internal (eyes/eyebrows, nose and mouth) and external (face contour, chin, ears and hair) features (Caldara et al., 2005; Le Grand et al., 2006). Behavioural experiments have suggested that both face regions provide diagnostic information for optimal processing of both familiar and unfamiliar faces (Brunas, Young, & Ellis, 1990; Campbell, Walker, & Baroncohen, 1995; Ellis, Shepherd, & Davies, 1979; Longmore, Liu, & Young, 2015; Megreya & Bindemann, 2009; Mondloch, Le Grand, & Maurer, 2002; Shepherd, Davies, & Ellis, 1981; Simion, Cassia, Turati, & Valenza, 2001; Young, Hay, Mcweeny, Flude, & Ellis, 1985).
Whereas that impaired performance in familiarity judgement and face naming tasks is present in both acquired (Busigny & Rossion, 2010a; Ramon, Busigny, Gosselin, & Rossion, 2017) and developmental (Behrmann, Avidan, Marotta, & Kimchi, 2005; Duchaine, Yovel, Butterworth, & Nakayama, 2006) prosopagnosia, low performance in unfamiliar face matching tasks has also been frequently observed in these two neuropsychological entities (Behrmann, Avidan, Marotta & Kimchi., 2005; Busigny & Rossion, 2010a, 2010b; Duchaine et al., 2006; Schmalzl, Palermo, Harris, & Coltheart, 2009). In the present ERP study we presented a face feature matching task (with realistic line drawings of faces) with which we evaluated how two adult individuals, one with acquired and another with developmental prosopagnosia, perceive both external and internal features in order to create new face representations in memory. The knowledge of neural markers that characterise the processing of unfamiliar faces in functionally different types of prosopagnosia might contribute to elucidate the brain mechanisms that underlie an ineffective formation of face representations, and, subsequently, to define the neural pattern that must accompany optimal mental operations for perceiving and memorising different physiognomies.
Event-related potentials (ERPs) are an excellent tool to unravel the temporal dynamics that characterises those brain subroutines accounting for face processing. Indeed, several ERPs have been identified as neurophysiological correlates of different stages concerning the formation and activation of neural representations of faces (cf. Olivares, Iglesias, Trujillo-Barreto & Valdés-Sosa, 2015; Schweinberger & Neumann, 2016; for reviews). Thus, the initial processing of faces as complex visual stimuli can be reflected in the early occipital P1 (peaking ~ 100–130 msec), which has been related to the detection of primary visual cues (for instance, contour), suggestive of the presence of stimuli resembling faces in a visual context (Linkenkaer-Hansen, Palva, Sams, Hietannen, Aronen & Ilmoniemi, 1998; Olivares & Iglesias, 2008). After initial face detection, the robust posterior temporal N170 (Bentin, Allison, Puce, Perez, & McCarthy, 1996; Rossion & Jacques, 2011) seems more sensitive to second–order relationships and the organisational characteristics of a face, prior to identification. In turn, in very posterior regions, the P2/P200 is modulated by the distance-to-norm in a norm-based, multi-dimensional face space (Wuttke & Schweinberger, 2019). The amplitude of this ERP, which is also face-sensitive, reflects the degree to which a face's second-order spatial configuration is similar to a current prototype (Schweinberger & Neumann, 2016). The transient activation of individual face representations is fairly associated with the N250r/N250 response, negative waveform peaking ~250 msec at posterior temporal sites (Schweinberger, Huddy, & Burton, 2004; Schweinberger, Pfutze, & Sommer, 1995). Moreover, in longer latencies, N400-like responses have been reported in face pair and face-feature matching tasks, suggesting the retrieval of diverse information associated to face and person identity (cf. Barrett & Rugg, 1989; Bentin & Deouell, 2000; Boehm & Sommer, 2005; Debruille, Pineda, & Renault, 1996; Eimer, 2000; Jemel, George, Olivares, Fiori, & Renault, 1999; Mnatsakanian & Tarkka, 2003; Olivares, Iglesias & Rodríguez-Holguin, 2003; Olivares, Lage-Castellanos, Bobes, & Iglesias, 2018; Paller et al., 2007; Wiese & Schweinberger, 2008). According to both neuroimaging data and source reconstruction analyses, all these responses are originated by the activity of neural “core” populations situated mainly in occipito-temporal cortical areas, including the so-called “ventral visual stream”, with some involvement of “extended” areas that are situated more rostrally (Gobbini & Haxby, 2007; Guntupalli, Wheeler, & Gobbini, 2017; Ishai, 2008; Young, Frühholz, & Schweinberger, 2020).
In order to investigate the relevance of both external and internal features in unfamiliar face perception, in a previous ERP study (Olivares & Iglesias, 2008) with young healthy people we simulated the formation of face representations using realistic drawings (from an Identikit gallery) of faces as stimuli. Participants performed a face-feature matching task with three stimuli per trial in which external features (E), internal features (I) and complete (target) faces were displayed in a step-wise mode, varying the sequence of presentation of primes by means of two different stimulus lists: E-I and I-E. The ERP pattern observed was characterised by a larger mismatch effect (i.e., greater negative amplitudes) ~300–500 msec around posterior-central regions (a sort of N400-like component), evoked by whole faces in the E-I, as compared to the I-E sequence. This effect suggested that the initial presentation of external features, followed by internal ones, facilitated the integration of facial features into new face representations and, consequently, allowed the detection of structural incongruences in “recently acquired” face recognition units. Another interesting result was that, whereas with E-I only initial external features elicited a sort of P300/P3 response, in I-E this response was evoked by both types of primes (i.e., external and internal features). Since P3 has been associated to selective attention and to subsequent memory processing (Kamp, Bader, & Mecklinger, 2017; Polich, 2007; Yonelinas, 2002; for reviews; Wiswede, Rüsseler, Hasselbach, & Münte, 2006), we interpreted that this pattern would probably reflect stimulus relevance and a greater allocation of neurocognitive resources for processing prime-related inputs in the I-E sequence.
Regarding the nature of the stimuli (realistic like line-drawings of faces) used in this matching task, it is worthy of mention that face-related research increasingly notes how artificial or computer-generated faces only partially tap face expertise (Crookes et al., 2015). Image information as reflectance or texture in natural pictures is thought to be relevant to efficiently discriminate different identities (Andrews, Baseler, Jenkins, Burton & Young., 2016; Itz, Schweinberger, & Kaufmann, 2017; Russell & Sinha, 2007). However, in face learning, both texture and shape seem relevant, as is evidenced by the finding that when new faces are shape- or texture-caricatured they are better memorized than their veridical counterparts (Itz et al., 2017). Moreover, shape seems more important for identity processing of unfamiliar faces as opposed to colour or reflectance information, which, in turn, seems more relevant for familiar face processing (Itz, Schweinberger, & Kaufmann, 2016). In the present study, which deals with unfamiliar faces, the use of artificial stimuli was motivated by the intention of precluding as much as possible the use of semantic associations that might interfere in processes thought to be exclusively of a visual-facial nature. Thus, we wished to replicate, as faithfully as possible, those conditions from our previous experiment with younger people.
In the present experiment, we applied the face-feature matching task to E.C. and I.P., with acquired and developmental prosopagnosia, respectively. The main aim here was to differentiate their brain responses concerning the visual processing of unfamiliar faces, from those of a group of age-matched controls. E.C. is a woman with damage in right occipito-temporal and parietal cortices. In the neuropsychological testing she exhibited a very poor performance in those visual tasks involving feature integration and configural face processing as well as a severe difficulty to learn new faces. In turn, I.P. showed a specific deficit in familiar face recognition and naming as well as in new face learning, with no apperceptive deficit. Since we considered that the magnitude of the mismatch effect evoked by targets ~300–500 msec in the face-feature matching task can be indicative of the efficiency in the integration of face components into face configurations, we searched firstly for modulations of this effect in the prosopagnosic participants, which would reveal the quality of formation of new face memories in these individuals, when compared to age-matched controls. Importantly, according to our previous findings, we expected to find some modulations in P3-like responses (as associated to primes in the face-feature matching task) in prosopagnosic individuals, since it might denote altered attentional mechanisms for efficiently encoding diagnostic face structural information, that might be crucial for an appropriate construction of face representations.
Modulations in face-sensitive ERPs in shorter latencies might be also expected, since they have been previously reported in impaired unfamiliar face perception (Alonso-Prieto, Caharel, Henson, & Rossion, 2011; Bentin, Deouell, & Soroker, 1999; Dalrymple et al., 2011; DeGutis, Bentin, Robertson, & D'Esposito, 2007; Eimer & McCarthy, 1999; Minnebusch, Suchan, Ramon, & Daum, 2007; Németh, Zimmer, Schweinberger, Vakli, & Kovács, 2014; Righart & de Gelder, 2007; Towler, Gosling, Duchaine, & Eimer, 2012) as indicative of disruption in very initial stages of face processing. Thus, we also hypothesised that altered neurocognitive mechanisms regarding the processing of new faces might be evidenced by amplitude decreases and topographic differences in face/identity-sensitive brain responses between prosopagnosic and typical individuals. Moreover, according to the functional meaning of face-sensitive ERPs, we might expect that E.C. would show a pattern of altered responses around earlier latencies (i.e., P1, N170), reflecting her individual perceptual difficulties related to face configuration processing, while I.P. would display a more preserved pattern in initial stages with modulations in later responses rather linked to more elaborate stages for face recognition.
In order to advance in the characterisation of the neural bases underlying inefficient unfamiliar face processing, we also conducted, in the present study, source reconstruction of the mismatch effects (when observed) in targets (complete faces) with a Bayesian Model Averaging approach (“BMA”, Trujillo-Barreto, Aubert-Vázquez, & Penny, 2008; Trujillo-Barreto, Aubert-Vázquez & Valdes-Sosa, 2004). This is an innovative source analysis method, considering both its methodological approach and the type of derived results. BMA is based on the Bayesian inference framework (MacKay, 1992) to offer inverse solutions of scalp-recorded ERPs, and it copes with one of the most common limitations of other existing source reconstruction methods, namely, model uncertainty, which arises when dealing with different solutions for a single data. BMA applies here the well-known LORETA method (Pascual-Marqui, 2002; Pascual-Marqui, Michel, & Lehmann, 1994) for source reconstruction and incorporates a further level of inference, that is, Bayesian model averaging. BMA belongs to the so-called “sparse” solutions, which are growing increasingly popular in the literature since they offer inverse solutions ranging from dipolar activations that include a few voxels to more distributed clustered activations. In addition, this methodology addresses two of the main problems that characterise linear inverse solutions, namely, the existence of ghost sources and the tendency to underestimate deep activity. Moreover, the present source reconstruction approach has proved good source localisation in face processing ERP studies (Bobes, Fernandez, Lopera, Quiroz, Galan, Vega, Trujillo, Valdes-Sosa & Valdes-Sosa, 2010; Bobes, Lage-Castellanos, Olivares, Hidalgo-Gato, Iglesias, Castro-Laguardia & Valdes-Sosa, 2019; Olivares et al., 2018; Olivares, Saavedra, Trujillo-Barreto, & Iglesias, 2013).
Section snippets
Prosopagnosic participants
E.C. (right handed, 30 y.o., University level of education) is an acquired visual agnosic woman, who suffered from a herpetic meningoencephalitis at the age of 13 and, as a consequence, her right occipito-temporal and parietal cortices (encompassing the ventral visual stream) were damaged. CT-scan images of E.C's lesion are included in Supplementary Material 1_CT-scan. She was diagnosed of left homonymous hemianopsia, with remaining ocular functions preserved and corrected visual acuity. Her
Behavioral results
Table 1 shows the average probabilities of hits (correct responses to targets in the match condition) and false alarms (incorrect responses to targets in the mismatch condition) in each sequence of presentation of features (E-I, I-E), as well as their corresponding discrimination sensitivity (d’) and bias (log β) values across participants. These behavioral variables indicate that the two types of presentation were easily performed by Controls and, according to their (negative) log β, they all
Discussion
In the present ERP experiment we aimed to explore how structural information related to unfamiliar faces is perceived by two individuals with prosopagnosia (one acquired and one developmental) in order to potentially create new face representations in memory. Thus, we contrasted the evoked neurophysiological activity of these prosopagnosic participants when performing a face-feature matching task regarding new faces, with that of a control group of typical adults, in order to delineate those
Declaration of competing interest
None.
Acknowledgments
The present study was supported by “Ministerio de Ciencia, Innovación y Universidades de España” (Grant PGC2018-094937-B-100). We are grateful to Brad Duchaine, who facilitated us both the first contact with patient I.P. and a face detection test. We also thank Andrea Álvarez-San Millán for her help in data acquisition from the Control group and edition of both Fig. 1 and Supplementary Material 3_Stimuli.
References (118)
- et al.
Working memory circuit as a function of increasing age in healthy adolescence: A systematic review and meta-analyses
NeuroImage: Clinical
(2016) - et al.
Contributions of feature shapes and surface cues to the recognition and neural re-presentation of facial identity
Cortex; a Journal Devoted To the Study of the Nervous System and Behavior
(2016) - et al.
Event-related potentials and the semantic matching of faces
Neuropsychologia
(1989) - et al.
Congenital prosopagnosia: Face-blind from birth
Trends in Cognate Science
(2005) - et al.
Neural correlates of intentional and incidental recognition of famous faces
Cancer Biotherapy & Radiopharmaceuticals
(2005) Measuring handedness with questionnaires
Neuropsychologia
(1977)- et al.
Acquired prosopagnosia abolishes the face inversion effect
Cortex; a Journal Devoted To the Study of the Nervous System and Behavior
(2010) - et al.
The development of differential use of inner and outer face-features in familiar face identification
Journal of Experimental Child Psychology
(1995) - et al.
Investigation of the single case in neuropsychology: Confidence limits on the abnormality of test scores and test score differences
Neuropsychologia
(2002) - et al.
The anatomic basis of the right face-selective N 170 in acquired prosopagnosia, a combined ERP/fMRI study
Neuropsychologia
(2011)
N400-like potentials elicited by faces and knowledge inhibition
Cancer Biotherapy & Radiopharmaceuticals
The Cambridge face memory test: Results for neurologically intact individuals and an investigation of its validity using inverted face stimuli and prosopagnosic participants
Neuropsychologia
Event-related brain potentials distinguish processing stages involved in face perception and recognition
Clinical Neurophysiology: Official Journal of the International Federation of Clinical Neurophysiology
Thresholding of statistical maps in functional neuroimaging using the false discovery rate
Neuroimage
A new method for off-line removal of ocular artifact
E Clinical Neurophysiology
Let's face it: it's a cortical network
Neuroimage
What aspects of face processing are impaired in developmental prosopagnosia?
Brain and Cognition
Face selective processing in human extrastriate cortex around 120 ms after stimulus onset revealed by magneto- and electroencephalography
Neuroscience Letters
The many faces of configural processing
Trends in Cognate Science
A probabilistic atlas of the human brain: Theory and rationale for its development
Neuroimage
Matching of familiar faces and abstract patterns: Behavioral and high-resolution ERP study
International Journal of Psychophysiology
Neuropsychological impairments in the recognition of faces, voices, and personal names
Brain and Cognition
The assessment and analysis of handedness: The Edinburgh handedness inventory
Neuropsychologia
Brain potentials and integration of external and internal features into face representations
International Journal of Psychophysiology
Making features relevant: Learning faces and event-related potentials recording using an analytic procedure
British Journal of Psychology
Long-term information and distributed neural activation are relevant for the ‘‘internal features advantage’’ in face processing: Electrophysiological and source reconstruction evidence
Cortex; a Journal Devoted To the Study of the Nervous System and Behavior
The activation of visual face memory and explicit face recognition are delayed in developmental prosopagnosia
Neuropsychologia
Low resolution electromagnetic tomography: A new method for localizing electrical activity in the brain
International Journal of Psychophysiology
Updating P300: An integrative theory of P3a and P3b
Clinical Neurophysiology: Official Journal of the International Federation of Clinical Neurophysiology
Modelling neural correlates of working memory: A coordinate based meta-analysis
Neuroimage
Repetition effects in human ERPs to faces
Cortex; a Journal Devoted To the Study of the Nervous System and Behavior
Early (N170/M170) face-sensitivity despite right lateral occipital brain damage in acquired prosopagnosia
Frontiers in Aging Neuroscience
Structural changes accompanying memory storage
Annual Review of Physiology
Event-related potentials and the matching of familiar and unfamiliar faces
Neuropsychologia
Detailed exploration of face-related processing in congenital prosopagnosia. 1: Behavioral findings
Journal of Cognition Neuroscience
Electrophysiological studies of face perception in humans
Journal of Cognition Neuroscience
Structural encoding and identification in face processing: ERP evidence for separate mechanisms
Cognitive Neuropsychology
Selective visual streaming in face recognition, evidence from developmental prosopagnosia
Neuroreport
The hippocampus and memory: Insights from spatial processing
Nature Reviews. Neuroscience
ERP generator anomalies in presymptomatic carriers of the Alzheimer's disease E280A PSI mutation
Human Brain Mapping
ERP source analysis guided by fMRI during familiar face processing
Brain Topography
Recognizing faces
Repetition priming from incomplete faces -evidence for part to whole completion
British Journal of Psychology
Acquired prosopagnosia is not due to a general impairment in fine-grained recognition of exemplars of a visually homogeneous category
Behavioural Neurology
Does prosopagnosia take the eyes out of face representations? Evidence for a defect in representing diagnostic facial information following brain damage
Journal of Cognition Neuroscience
Automatic 3D intersubject registration of MR volumetric in standardized Talairach space
J. Comput. Assist. Tomogr
Point and interval estimates of effect sizes for the case-controls design in neuropsychology: Rationale, methods, implementations, and proposed reporting standards
Cognitive Neuropsychology
Comparing an individual's test score against norms derived from small samples
The Clin Neuropsychologist
How well do computer-generated faces tap face expertise?
Plos One
Prosopagnosia: Anatomic basis and behavioral mechanisms
Neurology
Cited by (1)
The Neural Correlates of Developmental Prosopagnosia: Twenty-Five Years on
2023, Brain Sciences