A family at risk: Congenital prosopagnosia, poor face recognitionand visuoperceptual deficits within one family

Andreas Johnen a,b,n, Stefan C. Schmukle c, Judith Hüttenbrink a, Claudia Kischka d,Ingo Kennerknecht d, Christian Dobel a

a Institute for Biomagnetism and Biosignalanalysis, Westfälische Wilhelms-Universität Münster, Münster, Germanyb Department of Neurology, University Hospital Münster, Münster, Germanyc Department of Psychology, University of Leipzig, Leipzig, Germanyd Institute for Human Genetics, Westfälische Wilhelms-Universität Münster, Münster, Germany

a r t i c l e i n f o

Article history:Received 25 May 2013Received in revised form21 March 2014Accepted 25 March 2014Available online 3 April 2014

Keywords:Face processingProsopagnosiaDevelopmentalCongenitalFamily studyPerceptual organization

a b s t r a c t

Congenital prosopagnosia (CP) describes a severe face processing impairment despite intact early visionand in the absence of overt brain damage. CP is assumed to be present from birth and often transmittedwithin families. Previous studies reported conflicting findings regarding associated deficits in nonfacevisuoperceptual tasks. However, diagnostic criteria for CP significantly differed between studies,impeding conclusions on the heterogeneity of the impairment. Following current suggestions for clinicaldiagnoses of CP, we administered standardized tests for face processing, a self-report questionnaire andgeneral visual processing tests to an extended family (N¼28), in which many members reporteddifficulties with face recognition. This allowed us to assess the degree of heterogeneity of the deficitwithin a large sample of suspected CPs of similar genetic and environmental background. (a) We foundevidence for a severe face processing deficit but intact nonface visuoperceptual skills in three familymembers – a father and his two sons – who fulfilled conservative criteria for a CP diagnosis onstandardized tests and a self-report questionnaire, thus corroborating findings of familial transmissionsof CP. (b) Face processing performance of the remaining family members was also significantly below themean of the general population, suggesting that face processing impairments are transmitted as acontinuous trait rather than in a dichotomous all-or-nothing fashion. (c) Self-rating scores of facerecognition showed acceptable correlations with standardized tests, suggesting this method as a viablescreening procedure for CP diagnoses. (d) Finally, some family members revealed severe impairments ingeneral visual processing and nonface visual memory tasks either in conjunction with face perceptiondeficits or as an isolated impairment. This finding may indicate an elevated risk for more generalvisuoperceptual deficits in families with prosopagnosic members.

& 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Individual recognition of familiar faces is one of the mostimportant and demanding abilities for humans in social life (e.g.,Farah, Wilson, Drain, & Tanaka, 1998; Young, De Haan, & Bauer, 2008).The very high performance in this skill is assumed to be subserved bycortical networks specialized on the processing of faces (e.g., Haxby,Hoffman, & Gobbini, 2000; Kanwisher, McDermott, & Chun, 1997).Lesions within these cortical networks can lead to a state in whichpatients are dramatically impaired in recognizing faces, despitenormal lower-level vision, object identification skills, and semantic

knowledge. This severe neurological impairment has been calledprosopagnosia or face blindness and has attracted a lot of interest inthe last decades, both in the scientific community and in the generalpopulation. Prosopagnosia provides evidence that face processingis a cognitive function that may be dissociated from general visualprocessing or object processing (e.g., Farah, 1996; Moscovitch,Winocur, & Behrmann, 1997).

1.1. Congenital prosopagnosia

Individuals with an isolated face recognition deficit, whichmanifests itself in early childhood but is not attributable to overtneurological, neuropsychological, or psychiatric abnormalities,have been categorized as congenital, developmental, or hereditaryprosopagnosics (e.g., Behrmann & Avidan, 2005; Duchaine &Nakayama, 2006b; Jones & Tranel, 2001; Kennerknecht, Grüter,

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/neuropsychologia

Neuropsychologia

http://dx.doi.org/10.1016/j.neuropsychologia.2014.03.0130028-3932/& 2014 Elsevier Ltd. All rights reserved.

n Corresponding author at: Department of Neurology, University HospitalMünster, Albert-Schweitzer-Campus 1, 48149 Münster, Germany.Tel.: þ49 251 834 5304; fax: þ49 251 834 5313.

E-mail address: a.johnen@uni-muenster.de (A. Johnen).

Neuropsychologia 58 (2014) 52–63

Welling, & Wentzek, 2006; Kress & Daum, 2003). In line with ourearlier publications, we will use the term congenital prosopagno-sia (CP) to emphasize its presumed presence from birth and/or itshereditary origin. A significant number of case descriptions, butalso group- and family studies on this condition have been publi-shed in recent years (e.g., Dobel, Bol̈te, Aicher, & Schweinberger,2007; Duchaine, Germine, & Nakayama, 2007a; Grueter et al., 2007;Kennerknecht, Ho, & Wong, 2008a; Kennerknecht, Pluempe, &Welling, 2008b; Kennerknecht, Pluempe, Edwards, & Raman,2007; Schmalzl, Palermo, & Coltheart, 2008).

Independent research groups have estimated the prevalence of CPat 2–3% in the general population (Bowles et al., 2009; Kennerknechtet al., 2006, 2008a). However, it is not clear whether these findingsindicate a dichotomous, bimodal distribution of face processing skillsin the population, or whether CPs reflect the lower end of a normaldistribution, which would imply a continuous representation of faceprocessing skills in the general population. Recent evidence of so-called super recognizers who perform approximately 2 SDs above themean of the general population, but also population-based assess-ments with sensitive behavioral tests point towards the latterinterpretation of the results (Kennerknecht, Kischka, Stemper, Elze,& Stollhoff, 2011; Russell, Duchaine, & Nakayama, 2009; Wilmer,Germine, Chabris, Gerbasi, & Nakayama, 2012).

Given that many CPs report on first-degree relatives who arealso impaired in face recognition, most researchers argue for ahereditary contribution to CP (Behrmann & Avidan, 2005; DeHaan, 1999; Dobel et al., 2007; Galaburda & Duchaine, 2003;Kennerknecht et al., 2008b). Support for a genetic contribution toface recognition skills in the general population arises from studiesin behavioral genetics (Wilmer et al., 2010; Zhu et al., 2010). Thesestudies have compared performance of mono- and dizygotic twinson different tests of face cognition and have estimated the specificimpact of genetic variation in face recognition to be as high as 39%.Such findings imply that CP and face recognition deficits moregenerally may be predominantly found in certain families.

Regarding the underlying cognitive mechanism of the impair-ment, there is some evidence that persons with CP display abnorm-alities in what is called configural or holistic processing of faces (e.g,Avidan, Tanzer, & Behrmann, 2011; Palermo et al., 2011; Robbins &McKone, 2007; Tanaka & Farah, 1993; Van Belle, De Graef, Verfaillie,Busigny, & Rossion, 2010). Usually people perceive an upright face asan indecomposable whole, despite its constitution of individualfeatures with complex spatial relations to each other (Maurer,Grand, & Mondloch, 2002). Behavioral evidence for this specialcognitive treatment of upright faces comes most prominently, fromthe face inversion effect (Yin, 1969). The face inversion effect describesreduced recognition rates for faces that are presented upside down(i.e., inverted) compared to upright faces. This disproportion isconsiderably larger for faces compared with other objects. Suppo-sedly, the inversion of a face as well as a misalignment of the bottomhalf (composite face effect, e.g., Young, Hellawell, & Hay, 1987)interferes with the interactive processing of its parts, leading to afeature-based, analytic encoding strategy, which is less efficient thana holistic approach regarding accurate and fast recognition. Inprosopagnosic subjects however, these usually robust behavioraleffects are often not found; in fact, many described cases evendisplay better recognition rates for inverted faces (Avidan et al., 2011;Busigny & Rossion, 2011; Dobel, Putsche, Zwitserlood, & Junghof̈er,2008; Duchaine, Yovel, & Nakayama, 2007b; Farah, Wilson, Drain, &Tanaka, 1995; Lee, Duchaine, Wilson, & Nakayama, 2010; Schmalzl,Palermo, Harris, & Coltheart, 2009).

1.2. Heterogeneity of CP as a clinical condition

Whereas an impairment of face recognition is by definition atthe core of CP, evidence on associated neuropsychological deficits

in persons classified as CP is scattered and often conflicting,suggesting that CP may be a heterogeneous clinical condition(Dobel et al., 2007; Le Grand et al., 2006; Schmalzl et al., 2008).Among the reported perceptual deficits in nonface visual domainsare intraclass object agnosia (i.e., difficulties in discriminatingbetween members of other semantic categories such as housesor cars; Behrmann, Avidan, Marotta, & Kimchi, 2005; Duchaine etal., 2007b), impaired perception of biological motion including lipreading (Dobel et al., 2007; Lange et al., 2009), and visual imagerydeficits (Tree & Wilkie, 2010). In other cases, however, the facerecognition deficit was reported to be isolated, or at least notassociated with deficits in object processing or domain-generalvisual abilities (e.g., Duchaine & Nakayama, 2005; Stollhoff, Jost,Elze, & Kennerknecht, 2011).

One cause of this heterogeneity regarding associated deficitsmay be genuine diversity in the investigated cases themselves (i.e.,subjects may present with various subtypes of CP depending ongenetic and/or environmental factors). A second cause for thereported heterogeneity may be a poor comparability between theemployed methods and diagnostic criteria to classify individuals asCP. Such methods range from self-reports (e.g., Dinkelacker et al.,2010; Grueter et al., 2007; Kennerknecht et al., 2006) to inter-preting significant group differences in tailor-made experimentaltasks (e.g., Behrmann et al., 2005; Dobel et al., 2007; Duchaine& Nakayama, 2005; Le Grand et al., 2006). Moreover, amongthe latter types of studies the tested domains and neuropsycho-logical functions vary considerably in their level of specificity anddifficulty, ranging from low-level face perception of gender oremotion to highly abstract and difficult tasks on nonface percep-tual organization. From a practitioner's point of view, basicresearch on CP has suffered from a lack of consensus on clearcriteria both for diagnosis and for exclusion (Gainotti, 2010;Herzmann & Danthiir, 2008). However, the situation has beenimproved by Bowles et al. (2009), who suggested consensualclinical diagnostic criteria based on neuropsychological face pro-cessing tests which provide normative data, cut-off scores and ahigh level of psychometric quality.

With this study, we attempt to contribute to the current debateon heterogeneity of CP as a clinical entity. We analyzed the patternsof performance on face processing tasks and general visual proces-sing tasks across a large family sample (N¼28) in which manymembers reported face recognition difficulties. The studied sampleis highly similar with regard to genetic and environmental factors,especially within core families (e.g., a father and his offspring).This high group homogeneity allows us to largely control for theimpact of genetic and environmental variability on performance.Previous work on CP within family samples concluded that CP is aprimarily heterogeneous condition regarding associated neuropsy-chological deficits and/or underlying cognitive functions (Lee et al.,2010; Schmalzl et al., 2008). We were now interested in whetherthe use of the recently suggested standardized diagnostic criteriaand clinical cut-off scores (Bowles et al., 2009) might yield a morehomogeneous picture of the condition within families by reducingthe possibility of falsely diagnosing CP e.g., in ambiguous cases.For single-case diagnoses, such a normative account has severaladvantages over a group comparison of tailor-made experimentaltasks. First of all, normative samples are usually larger, allowingfor a more precise quantitative classification of the results comparedto smaller control group samples. Second, standardized taskshave usually been tested for their psychometric quality. Mostimportantly, published and standardized neuropsychological testsprovide a basis for reproducible results and comparisons betweenstudies and thus constitute the standard procedure in clinicalsettings.

With this normative account, we furthermore aimed to con-tribute to the current debates on whether CP is a categorical or a

A. Johnen et al. / Neuropsychologia 58 (2014) 52–63 53

continuous phenomenon as well as comment on the relevance andpracticability of self-reports in CP research.

1.3. Diagnostic criteria for CP

As recommended by Bowles et al. (2009), we employed twotests of face processing abilities for the diagnosis of CP: TheCambridge Face Memory Test (CFMT; Duchaine & Nakayama,2006b) and the Cambridge Face Perception Test (CFPT; Duchaineet al., 2007b). These computer-based tests have overcome variousshortcomings of older tests, such as feature matching of facial andnonfacial information (Duchaine & Weidenfeld, 2003; Duchaine &Nakayama, 2004) by using unfamiliar natural faces without hair-lines as stimuli. Recent studies from different laboratories on thepsychometric qualities of the CFMT establish this test as a reliableinstrument for assessing face memory and even its subtle impair-ments on an individual basis (Bowles et al., 2009; Herzmann &Danthiir, 2008; Russell et al., 2009). The CFPT moreover provides ameasure of a perceptual face inversion effect (i.e., the hallmark ofholistic face processing) by contrasting performance in trials withupright and inverted faces (Duchaine et al., 2007b). Despiteobvious differences between the two tasks, as well as recentevidence for dissociations between face perception and facememory abilities (Wilhelm et al., 2010), the CFMT and CFPT arestrongly correlated, and the authors of the tests suggest using bothfor a diagnosis of CP. In line with Bowles et al. (2009), we considera subject as clinically impaired if performance in both the CFMTand the CFPT is lower than 2 SDs below the predicted mean for theperson's age. We also employed a standardized self-report ques-tionnaire consisting of 15 items on everyday face processing skills,which has been used in previous studies (Kennerknecht et al.,2007; Kennerknecht et al., 2008a). The aim of this self-reportquestionnaire was to test the predicitve validity of such a quick-and-dirty screening method and compare it to the diagnoses basedon established neuropsychological tests. The major advantage of ascreening for prosopagnosic symptoms lies in its potential eco-nomic efficiency: Whereas in-depth neuropsychological testingmay ascertain a diagnosis, short screening questionnaires may beused to preselect subjects that report deficits. This can be done in afast and efficient way, e.g. via telephone or online. Such proce-dures have already proven to be useful for a range of otherneurocognitive syndromes including early dementia and develop-mental dyslexia (e.g. Ramlall, Chipps, Bhigjee, & Pillay, 2013;Snowling, Dawes, Nash, & Hulme, 2012).

Given the lack of clear guidelines for exclusion criteria, we decidedon a pragmatic procedure to screen for possible impairments that mayresult in face processing dysfunctions which cannot be attributed toCP. Since face recognition and/or face perception deficits are assumedto be common symptoms in several developmental (e.g., autismspectrum disorders; Weigelt, Koldewyn, & Kanwisher, 2012; Wilson,Freeman, Brock, Burton, & Palermo, 2010) and neurodegenerativedisorders (e.g., frontotemporal dementia; Omar, Rohrer, Hailstone,

& Warren, 2011), we specifically asked subjects for any birth complica-tions, developmental abnormalities as well as their neurological andpsychiatric history. To screen for major impairments in basic visualintegrity and visuoperceptual abilities which may impede a reasonableevaluation of face processing abilities, we employed two widely usedneuropsychological assessment tools. First, we administered the VisualObject and Space Perception Battery (VOSP; Warrington & James,1992) to preclude the possibility of a general visual agnosia as a causefor prosopagnosic symptoms. Following diagnostic recommendationsby Von Cramon, Mai, and Ziegler (1995), the VOSP combines the mostimportant tests for the diagnosis of visual agnosias. Second, weemployed the Rey–Osterrieth Complex Figure Test (RCFT; Meyers &Meyers, 1995) to screen for impairments in perceptual organizationand in figural memory for nonface material. Analogously to the faceprocessing tests, we only regard major impairments (i.e., an age-corrected score more than 2SDs below the norm) in either the RCFT orthe VOSP as an exclusion criterion for a CP classification.

1.4. Goals of the current study

In this study we aim to investigate familial patterns of faceand nonface processing deficits in an extended family sample ofsuspected CPs. We focus on standardized neuropsychologicalassessment and the use of operational diagnostic criteria, in orderto reduce the risk of falsely diagnosing CP. We are specificallyinterested in the following questions:

(a) Will we find similar neuropsychological profiles for tests offace and nonface visual processing in close relatives suggestive ofCP? This question relates to the aspect of homogeneity vs.heterogeneity of CP within a family sample. (b) Will those familymembers who are not suspected of having CP perform normally inface processing tasks compared to the general population? Thisaddresses the question of whether deficient face processing istransmitted in a dichotomous all-or-nothing or in a continuousfashion. (c) How are scores on standardized face processing testsreflected in self-report data? This addresses the question whetherself-report data can be used as a viable screening instrument toclassify persons with face processing deficits.

2. Material and methods

2.1. Participants

Twenty-eight members of a family, ranging from 13 to 55 years of age andstemming from two generations, participated in the study (see Fig. 1). Here, we willpresent data for the ten siblings and their offsprings. The spouses of the siblingswere screened with a semi-structured interview, but no indication for an impair-ment of face or nonface visual processing was found. The family sample can befurther subdivided into nine core families, i.e., nine of the ten siblings of the oldergeneration with their respective children (M45 was childless).

Preceding the neuropsychological assessment, a short oral interview wasconducted to assess level of education, occupation, neurological and psychiatrichistory, and visual acuity. All subjects reported at least average education levels:Most of the older subjects had attained the German “Mittlere Reife” (10 years of

Fig. 1. Family pedigree. Subject codes represent gender and age. Circles and squares represent female and male subjects, respectively. Crossed out subjects are deceased.Subjects with the label “n.t.” were not tested.

A. Johnen et al. / Neuropsychologia 58 (2014) 52–6354

general schooling), except for F53 and M52 who had a German “Hauptschulabs-chluss” (9 years of general schooling). All subjects were highly functioning ineveryday life with diverse occupations such as architectural draftsman, carpenter,saleswoman or gardener. Subjects from the younger generation were either still inschool (e.g., F15, M13), university students (e.g., M22, M21), or involved in jobtrainings and apprenticeships in multiple domains (e.g., M17, F17a), suggesting atleast average general intelligence levels. The 10 siblings all lived in single-familyhomes in small towns outside of Münster and close to each other. Most of theyounger generation lived at home with their parents, except for M23, M28, M22,and F33, who lived either on their own or with their spouses. Based on a grossclinical examination performed by IK, neurological status was confirmed to benormal for all subjects. Moreover none of the subjects reported any neurological orpsychiatric history, except for M21, who had recently experienced a single epilepticseizure and was consequently medicated with anticonvulsiva. All subjects deniedknowledge of any birth complications or early visual deprivations that could haveresulted in developmental face processing impairments (Le Grand, Mondloch,Maurer, & Brent, 2004). All subjects reported normal or corrected-to-normal vision.Before testing, subjects gave their written informed consent. Parents gave writtenconsent for minor children, after asking them whether they wanted to participatein the study or not. By request, instead of individual compensation, the family as awhole received a shared reimbursement of 500 € for their participation.

The study was carried out according to the ethical approval of the “Kommissionder Ärztekammer Westfalen-Lippe und der Medizinischen Fakultät derWestfälischen-Wilhelms Universität Münster (3XKenn2)”.

2.2. Self-rating questionnaire for face processing

In the self-rating questionnaire subjects were presented with statementsconcerning face processing, like “It takes me a long time to recognize faces” or “Irecognize famous people immediately”. They were asked to indicate their agree-ment on a 5-point Likert scale (Cronbach's α¼ .79 for our sample; a translation of all15 statements can be found in Kennerknecht et al., 2007). Since this questionnairehas been used in previous studies, we were able to compare the results of oursubjects against a normative control sample (N¼115) (Kennerknecht et al., 2007).For this purpose, we first summed up the raw values for each subject across all 15statements. Higher values represent more reported deficits in face processing. Wenext computed z scores for each subject by subtracting the mean of the controlsample and dividing by the control sample's SD.

2.3. Neuropsychological assessment

Tests were carried out at the subjects' homes in quiet rooms by either the firstor the third author. All tests were presented as stated in the manuals, either as apaper-pencil version or on a laptop computer with a 38.2 cm screen. Subjects weregranted comfortable viewing and lighting conditions, with approximately 50 cmdistance from the computer screen. The order of tests was fixed for all subjects (i.e.,Self-Rating Questionnaire, CFMT, RCFT Copy trial, VOSP Silhouettes, RCFT Immedi-ate Recall, VOSP object decision, VOSP number location, VOSP cube analysis, CFPT,RCFT Delayed Recall, RCFT Recognition).

2.3.1. The Cambridge Face Memory Test (CFMT)To assess face recognition and memory for faces, we administered the upright

version of the Cambridge Face Memory Test (CFMT; Duchaine & Nakayama, 2006b).The CFMT consists of three stages, in which six target faces have to be recognizedunder increasingly challenging conditions. In the first stage, subjects are presentedwith three different views of six unfamiliar faces, one at a time. In each trial, thesubject is asked to memorize the face and identify it later from two additional facesby typing 1, 2, or 3 on a computer keyboard. In the two later stages, novel views andGaussian-blurred versions of the learned faces have to be identified, increasing thedifficulty of the task. Raw scores for each subject were derived by summing thenumber of correct answers out of 72 trials. To correct for aging effects in facerecognition, age-corrected comparison scores were calculated for each subject,using a formula that represents the best fit curve for the relation between age andperformance in the normative data sample of the CFMT (Bowles et al., 2009). In afinal step, z-scores were calculated for each subject based on the age-correctednormative scores to facilitate comparison and clinical diagnoses. Internal consis-tency of the CFMT is high (Cronbach's α¼ .89) and sufficient for single-casediagnosis of face recognition deficits (Bowles et al., 2009; Herzmann & Danthiir,2008).

2.3.2. The Cambridge Face Perception Test (CFPT)The CFPT is a computer-based measure of facial-similarity judgment abilities.

It requires the subject to rank a set of six faces in descending order of similarity to asimultaneously presented target face (Duchaine et al., 2007b). The stimuli werecreated by morphing different faces into the target face to obtain sets of six faces,each with a different percentage of similarity to the target face. Each trial has a timelimit of 1 min. The CFPT consists of eight upright trials and eight inverted trials, inwhich the subject has to judge a set of inverted faces for their similarity to a target

face, which is also inverted. The sets are presented in a randomized order, and thesubject is asked to sort the images using a computer mouse. Error scores for theCFPT are derived by summing the deviations of all items from their correctpositions at the end of each trial. The difference between scores for upright andinverted trials on the CFPT can be used as a measure of a perceptual face inversioneffect (Duchaine et al., 2007a, 2007b; Russell et al., 2009). Note that a negative valuetherefore denotes the lack of a face inversion effect (i.e., inverted faces are processedas well as or better than upright faces). Similarly to the procedure adopted for theCFMT, individual test scores for the upright trials and the inversion effect measurewere compared to an age-corrected predicted score, based on the normative dataprovided by Bowles et al. (2009), and the raw scores were transformed to z-scores.Cronbach's α of the CFPT upright is.74, whereas the inverted condition reachesonly.50 (Bowles et al., 2009).

2.3.3. The Visual Object and Space Perception Battery (VOSP)To assess visual integrity as well as basic shape detection and object identifica-

tion abilities, we administered the German version of the VOSP (Warrington &James, 1992). The VOSP is a well-established neuropsychological test battery, whichis widely used in clinical settings to reliably identify deficits in object and spaceperception by comparing individual raw scores to scores of a neurologically healthynormative sample (Warrington & James, 1992). Due to expected ceiling effects,considerations of the total test duration, and functional overlaps with othersubtests (Bonello, Rapport, & Millis, 1997), we excluded the screening subtest andtwo subtests from the space perception part of the VOSP (Dot Counting and PositionDiscrimination). The stimuli of the five remaining subtests were presented asprinted cards, and untimed answers were given verbally. From the object percep-tion part of the test, we present results for the subtests Silhouettes (Cronbach'sα¼ .78) and Object Decision (α¼ .58; all reliabilities from Bonello et al., 1997), whichmeasure basic object identification skills by asking subjects to identify silhouettesof objects depicted from unusual angles. The ability to identify these degradedobjects depends strongly on efficient perceptual organization in order to merge thefeatural elements of the silhouettes into a fully structured global percept (Bonelloet al., 1997). We do not present results for the subtest Progressive Silhouettes due toits very low reliability (α¼ .27). From the space perception part of the battery, weadministered the subtests Number Location (α¼ .84) and Cube Analysis (α¼ .77).Both tests are sensitive to deficits in spatial perception measured by asking thesubject to discriminate between relative spatial positions or to interpret a three-dimensional cube represented in two dimensions (Bonello et al., 1997). Correctanswers were recorded and summed for each subtest and subject. These raw scoreswere then standardized by subtracting the mean of the respective normativesample and dividing by its standard deviation.

2.3.4. The Rey Complex Figure Test and Recognition Trial (RCFT)To assess visual memory functions for complex figural material as well as

perceptual organization skills, we administered the RCFT (Meyers & Meyers, 1995).The RCFT uses the well-known original Rey–Osterrieth complex figure (Rey, 1941)as stimulus material. The configural structure of the figure entails an organizationof its feature elements into a global representation. Similar to faces, the complexfigure has a configural structure (i.e., the whole is composed of smaller parts), andthe organizational encoding strategy for copying the figure is highly predictive ofthe performance in reproducing it from memory: With regard to memoryperformance on recall trials, organizing the figure from a global vantage pointand clustering it into meaningful global units is more effective than encodingisolated items in a piecemeal fashion (Akshoomoff & Stiles, 1995; Anderson,Anderson, & Garth, 2001; Deckersbach et al., 2000; Newman & Krikorian, 2001).Thus, recall trials of the RCFT measure individual efficiency of perceptual organiza-tion and strategy beyond figural memory (Kramer & Wells, 2004; Newman &Krikorian, 2001). The test was carried out according to the rules in the manual:Subjects were asked to copy the figure onto a blank sheet of paper (the copy trial),and then to reproduce it from memory after a 3-min delay (the immediate recall).After an additional 30-min interval, subjects were asked to reproduce the figureagain from memory (the delayed recall) followed by the recognition trial, for whichparts of the complex figure were shown together with distractor items. To preventmemory interference effects, we ensured that the intervals between trials did notcontain tasks that also tapped into memory functions or used perceptually similarstimuli. Scoring of the RCFT trials was carried out by the first author: Based on thestandard scoring system by Rey (1941), presented in Lezak, Howieson, Bigler, andTranel (2012), the figure was divided into 18 different units. For the copy trial andboth recall trials, each unit was rated separately for accuracy and placement. Therecognition trial raw score was calculated by adding hits and correct rejectionstogether. Raw scores were compared to normative data based on a sample of 601healthy adults (18–89 years) and a sample of 505 normal children and adolescents,aged 6 to 17, and transformed into age-corrected z scores (Meyers & Meyers, 1995).Temporal stability over an average retest interval of 184 days is rtt¼ .76 for theimmediate recall, rtt¼ .89 for the delayed recall, and rtt¼ .87 for the recognition trial,respectively (Meyers & Meyers, 1995).

A. Johnen et al. / Neuropsychologia 58 (2014) 52–63 55

2.4. Data analysis

Statistical analyses were carried out using SPSS 21. Age-corrected z scores werecalculated for each neuropsychological test. As a general principle, we will interpretz scores below �2 as “impaired” or “deficient”, and z scores between �1.5 and �2as “marginally impaired”. We additionally calculated z scores to test the signifi-cance of differences at an individual level as suggested by Payne & Jones (1957).These z scores take into account the reliabilities of the used tests and thus provideinformation on the confidence that the observed difference is not merely due tomeasurement errors. We compared both face tasks (CFMT & CFPT) against eachother and also against those nonface screening tests, in which we discuss deficits insome subjects. We conservatively decided to report the two nonface screeningmeasures with the lowest reliabilities (RCFT Immediate recall & VOSP ObjectDecision) for this analysis.1

3. Results

Table 1 summarizes the results of all 28 subjects on the 12neuropsychological measures and the self-report.

3.1. Performance in face processing tasks and CP classifications

z Scores for the CFMT as well as the upright version of the CFPTand the perceptual face inversion effect measure are shown in thefirst three columns of Table 1. Three first-degree relatives – a fatherand his two sons (M46, M16b, M13) – each scored lower thanz¼�2 on both the CFMT and the upright CFPT, indicating a clear CPclassification. Additionally, these three subjects displayed z valuesaround �2 on the CFPT inversion effect measure, indicating thatthey did not show the typical perceptual face inversion effect. In fact,for M13 the face inversion effect was itself inverted—he made moreerrors on the upright than on the inverted trials of the CFPT (rawscores: CFPT upright¼64; CFPT inverted¼58). None of these threesubjects showed significant deficits on any measures of the VOSP orthe RCFT, with all z scores higher (i.e., better performance) than �2.M16b displayed a single marginal z score of �1.63 in the objectidentification subtest of the VOSP. Table 2 furthermore demonstratesthat all three CPs display significant differences between face andnonface tasks, indicating that the observed discrepancies in thesesubjects are highly unlikely to be merely due to measurement errorsof the employed tasks.

Besides this homogeneous group of three first-degree relativeswith an impairment in face processing, another four subjects(M23, M16a, M27, M21; note that M27 and M21 are brothers)displayed a very similar pattern of results on face tasks. They allshowed impaired or at least marginally impaired performance inthe CFMT and in at least one measure of the CFPT, though withoutreaching the conservative diagnosis cut-off of z¼�2.00 on bothtests. Of these four subjects, and in contrast to the CP triplet, M21and M23 also revealed deficient scores on two trials of the RCFT,pointing towards deficits in figural memory functions and generalperceptual organization skills. We will refer to this group ofsubjects as bad recognizers.

Finally, a pattern of normal face recognition (CFMT) butdeficient facial similarity judgment abilities (CFPT upright) wasfound in a number of subjects (F48, F23, M50, M49, M18; Table 1).For two of these subjects (F48, M18), this discrepancy betweenCFMT and CFPT upright was found to be statistically significant(Table 2, column 5). While these subjects are certainly somewhatlimited in their ability to perceptually judge similarities of faces,their face recognition ability – the core feature of CP – is spared.This important issue of partly dissociated skills underlying the

CFMT and CFPT will again be addressed in a correlational analysisin Section 3.5.

Given that all three classified CP and all of the bad recognizerswere males, we have further investigated the gender distribution offace recognition deficits in our data despite the fact that we had no apriori hypotheses concerning this aspect.2 The percentage of subjectsthat were classified as CP did not significantly differ by gender, χ2(1,N¼28)¼1.86, p¼ .17. However there was a significant gender effectwhen CP and bad recognizers were both considered as impaired, χ2(1,N¼28)¼5.18, p¼ .03 suggesting that male subjects were more likelyto be classified as CP or bad recognizers in our sample.

3.2. Face processing of the family sample as a group

We next compared the family's group mean against the mean ofthe general population, taken from available normative data (Bowleset al., 2009). As can be seen in the bottom row of Table 1, the resultindicates a significantly lower group mean for both face processingtests, compared with the general population mean—CFMT (M¼� .90,SD¼ .87), t(27)¼�5.51, po.001; CFPT upright (M¼�1.20, SD¼1.30),t(27)¼�4.94, po.001). When the three CPs (M46, M16b, M13) aswell as the group of bad recognizers who failed to reach all diagnosticcriteria (M23, M16a, M27, M21) were taken out of this analysis, thegroup means of the remaining subjects on the CFMT and the CFPTupright still remained significantly below average—CFMT, (M¼� .55,SD¼ .67), t(20)¼�3.70, po.001; CFPT upright, (M¼� .67, SD¼1.00), t(20)¼�3.06, p¼ .006. In order to rule out the possibility that theseremaining subjects may suffer from more general cognitive impair-ments, which globally affect their performance on both face as well asnonface tasks, we furthermore directly compared mean performanceon face tasks vs. nonface tasks in this group. Face tasks (CFMT, CFPTupright) and nonface tasks respectively were aggregated by summingup z scores and dividing by the number of tasks. A paired t-testrevealed that for the group of subjects who were neither classified asCP nor as bad recognizers (i.e., remaining subjects), scores on theaggregated face tasks were significantly lower than on the aggregatednonface tasks, t(20)¼3.13, p¼ .005.

Thus, the family as a whole displayed a tendency to performbelow average on both face processing tasks, even without thoseindividuals classified as CP or as bad recognizers.

3.3. Comparison of self-report data and results in face processingtests

In order to compare the self-report data with the neuropsy-chological test data for face processing, we adopted a twofoldstrategy—a single-subject and a group analysis. Regarding indivi-dual self-reports of the classified CPs, subjects M46 and M16bconsider their ability to recognize familiar persons as severelyimpaired, with z scores of �2.12 and �3.4, respectively. Thisroughly matches their individual performances in the CFMT andCFPT. M13 judges himself less impaired, but his self-rating is alsomarginally impaired, z¼�1.58. Of the four subjects who failed toreach all diagnostic criteria and were thus classified as badrecognizers, only M27 rates his own face processing abilities asmarginally impaired, z¼�1.58. In the remaining three subjectsthe somewhat conspicuous findings in the CFMT and CFPT are notclearly reflected in their self-reports. For the rest of the family, theself-reports vary considerably, with values ranging from z¼ .79(M18) to z¼�4.31 (M52). Just as with the neuropsychological testdata, the mean self-rating of the family as a group is significantlylower than in the available normative sample, (M¼�1.07, SD¼1.30), t(27)¼�4.53, po .001. Also analogously to the results in1 The statistical procedure was kindly suggested by an anonymous reviewer. All

other possible test comparisons can be calculated with the data provided in Table 1and the reliabilities reported in the methods section, and are in line with ourinterpretations. 2 We owe this observation to an anonymous reviewer.

A. Johnen et al. / Neuropsychologia 58 (2014) 52–6356

standardized test data, this even holds true when the three CPsand the four subjects, which we categorized as bad recognizers areexcluded from this analysis, (M¼� .94, SD¼1.4), t(21)¼�3.07,

p¼ .006). Thus, the general tendency of this family to pe-rform below average in face processing tests is mirrored in self-report data.

Table 1Individual z scores for all test variables.

CFMTtotal

CFPT upright

CFPT inversion

effect

Self-Rating

RCFT copy trial

RCFT immediate

recall

RCFT delayed recall

RCFT recognition

VOSP silhouettes

VOSP object

decision

VOSP number location

VOSP cube

analysisF55 -1.60 -.75 .80 -2.31 -.39 -1.23 -1.24 -2.02 -1.30 -3.00 .55 .67

M28 -.08 -1.46 -1.70 -2.67 -.43 -2.59 -2.92 -2.14 .46 -1.00 .55 .58F23 -.87 -1.77 .25 .24 -1.53 -2.48 -2.66 -.96 -.27 -1.00 -3.09 -.25

M54 -.10 -.14 .39 -.67 -.12 -1.07 -1.60 .31 -1.30 -1.95 -.36 -.17M23 -1.82 -2.76 -.72 -1.03 -1.53 -1.84 -1.12 2.00 -.27 .25 .55 .58

F53 -1.78 .14 -.03 -2.49 -.91 -.18 -1.23 -.28 -1.05 -.89 .55 -1.00F33 .03 .88 .80 -1.21 .43 1.49 1.56 -.46 .22 -.38 -2.18 .58M27 -1.85 -2.78 -1.70 -1.58 .70 -.42 -.08 -.32 -.02 -.38 -.36 .58M21 -1.67 -2.41 -2.67 -1.21 .52 -1.97 -2.02 1.26 -.02 -.38 -.36 -.25

M52 -1.22 .26 1.08 -4.31 -1.71 -.06 -.12 .31 -.55 -1.95 -2.18 .67F19a .13 .74 -.17 -.31 -.06 1.52 .83 .85 -.02 -1.00 .55 -1.08F15 -.98 -.01 -.72 -1.03 1.17 1.56 2.02 1.67 -.51 -2.25 -.36 .58

M50 -.92 -1.80 -.86 -.12 -2.50 -.95 -1.22 -1.46 -.55 -1.42 .55 -.17F19b -.46 -.91 -.72 .24 .50 -.39 -.45 -3.20 -.76 -2.88 .55 .58F17a .63 -.05 .94 .60 -.89 -.42 -.54 .75 .46 -.38 .55 -.25

M49 -.82 -1.84 -.58 -3.03 -.18 -.46 -.73 .07 .71 -1.00 .55 .58F17b .28 -.05 -1.28 -1.21 .24 1.56 1.34 -.53 -.27 -1.00 -.36 .58M16a -1.83 -2.34 -1.28 -1.21 -.39 -.05 -.11 -.90 -.27 -.38 -.36 .58

F48 .10 -2.70 -.72 -.85 -.50 -2.95 -2.54 -.92 -1.00 -3.50 .55 .58M18 -.09 -2.21 -1.97 .79 -.05 -.18 .40 -1.18 -1.24 -1.63 .55 .58M15 -.39 -.50 .94 -.12 -.62 .24 .34 .45 -.51 -1.00 .55 .58

M47 -1.22 -.27 .25 -2.12 -.18 .27 -.18 -.92 1.20 .87 -.36 -.25M22 .08 .38 -.45 .06 .53 -.81 -.10 1.26 .22 -.38 .55 .58M17 -1.37 -.71 .39 .60 -.89 -.22 -.16 .11 .46 -1.00 .55 -1.08

M46 -2.07 -2.28 -1.70 -2.12 .23 -.46 .0 .07 .46 -1.00 .55 .58M16b -2.30 -3.66 -1.70 -3.40 .87 1.11 1.45 .33 -1.24 -1.63 -.36 .58M13 -2.35 -2.43 -2.25 -1.58 .09 .54 1.18 .50 -1.24 -.38 .55 .58

M45 -.79 -1.33 -1.42 .24 -.60 .45 .35 -.42 -.27 -.38 .55 .58Mean (SD) -.90 (.87) -1.2 (1.30) -.60 (1.07) -1.07 (1.3) -.30 (.84) -.36 (1.24) -.34 (1.30) -.21 (1.18) -.30 (.68) -1.11 (.98) -.04 (.97) .24 (.57)t -5.51** -4.94** -2.97** -4.53** -1.84 -1.52 -1.39 -.92 -2.36* -5.96** -.21 2.25*

Note. z Scores for all 28 subjects on neuropsychological tests and self-rating for face processing. Subjects of the younger generation were indented to the right, beneath theirparent. Cell colors represent severity of deficits. Black cells represent values more than 2 SDs lower than the mean of the normative sample. Dark grey cells indicate valuesbetween �1.5 and �2.0. White cells represent scores within or above average (i.e., z4�1.5). Rows below the table contain mean values for each test and whether thesesignificantly deviate from the normative sample (expressed as t-, and corresponding p-values, *po .05, **po .01). CFMT: Cambridge Face Memory Test; CFPT: Cambridge FacePerception Test; RCFT: Rey-Complex Figure Test; VOSP: Visual Object And Space Battery.

Table 2Tests of significance between individual test scores in face and nonface tasks.

CFMT vs. RCFT Immediate

CFMT vs.VOSP Object Decision

CFPT upright vs.RCFT Immediate

CFPT upright vs.VOSP Object Decision

CFMT vs.CFPT upright1

F55 -.63 1.83 .68 2.65 -1.40M28 4.24 .28 1.60 -1.42 2.26F23 2.72 -.81 1.00 -1.80 1.48

M54 1.64 2.44 1.32 2.10 .07M23 .04 -3.56 -1.30 -4.28 1.55

F53 -2.70 -1.33 .46 1.16 -3.16F33 -2.47 -.29 -.86 .77 -1.40M27 -2.42 -2.87 -3.33 -3.66 1.52M21 .50 -2.63 -.63 -3.22 1.22

M52 -1.96 .90 .45 2.59 -2.43F19a -2.36 .57 -1.10 1.24 -1.01F15 -4.29 .50 -2.22 1.62 -1.59

M50 .06 .59 -1.20 -.56 1.45F19b -.12 1.94 -.73 1.17 .73F17a 1.78 .54 .52 -.36 1.12

M49 -.61 -.74 -1.96 -1.89 1.68F17b -2.16 .78 -2.28 .29 .54M16a -3.00 -2.84 -3.23 -3.13 .84

F48 5.15 3.43 .35 -.37 4.61M18 .15 .99 -2.87 -1.69 3.48M15 -1.06 -.14 -1.05 -.27 .19

M47 -2.52 -3.47 -.77 -1.91 -1.56M22 1.51 -.22 1.68 .16 -.48M17 -1.95 -1.49 -.69 -.51 -1.09

M46 -2.72 -2.45 -2.58 -2.42 .35M16b -5.76 -2.04 -6.74 -3.44 2.23M13 -4.88 -3.56 -4.21 -3.25 .14

M45 -2.10 -1.42 -2.51 -1.91 .89

Note. Values represent z scores for testing the significance of differences between two individual test scores (Payne & Jones, 1957). Values larger than z¼71.96 indicate, thatthe observed difference is highly unlikely to be due to measurement errors only (po .05). Algebraic sign of the z score indicates the direction of the difference. For columns 1–4thus, black cells represent a discrepancy of face over nonface tasks. Grey cells indicate a discrepancy in the opposite direction (nonface4face tasks).1We included this analysis to substantiate our observation of partly dissociable face processing abilities in several individuals. For the comparison CFMT vs. CFPT upright, blackcells indicate a significant discrepancy of CFMT over CFPT performance. Grey cells contrarily indicate a significantly worse performance in the CFPT compared to the CFMT.

A. Johnen et al. / Neuropsychologia 58 (2014) 52–63 57

Correlations between tests for face processing and the self-reports are displayed in Table 3. The CFMT is moderately correlatedwith self-reported impairment, r(26)¼� .5, p¼ .007, suggesting thatperformance in the CFMT is at least partly reflected in the self-report data. Considering the generally reduced means on the faceprocessing measures and the consequent restriction of variance inthis family, the strength of this correlation is impressive. Althoughwe were able to replicate the finding by Bowles et al. (2009) of astrong association between the CFMT and the upright CFPT, r(26)¼ .57, p¼ .002, neither the perceptual face inversion effectmeasure from the CFPT, nor the upright CFPT is reflected in theself-report data, indicated by low non-significant correlations. Thefact that the items in our self-report battery (Kennerknecht et al.,2007) mainly focus on face recognition abilities and not on per-ceptual facial similarity judgement, may explain this result. Wefurthermore calculated a receiver operating characteristic (ROC)curve to estimate sensitivity and specificity as well as the optimalcut-off score for the self-report questionnaire given the CP diagnosesbased on the neuropsychological tests. The self-report questionnaireused in the current study shows an area under the curve (AUC) of.83 (95% confidence interval: .66–.99) for the distinction between CPand non-CP subjects. The optimal cut-off score for the self-reportquestionnaire as a first screening procedure was 34 in the currentsample with a sensitivity of 100% and a specificity of 72%. Thus, withthis score it is possible to identify all CPs with the self-reportquestionnaire, however at the expense of a substantial rate of falsepositives. These false positives are then classified by using neuro-psychological tests.

In summary, impaired face processing skills and self-reportedimpairment show a considerable overlap in this family. All threeclassified CPs reported at least subtle impairments in recognizingfaces. Although in some subjects the results diverged stronglybetween assessed face recognition and self-report (i.e., M49, M52),the high correlation between CFMT and self-reported face recog-nition deficits as well as the accurate classification parametersindicated by the ROC analysis suggests that self-report data can bea useful screening tool for CP classifications.

3.4. General visuoperceptual deficits

Even though we had no a priori hypotheses about generalvisuoperceptual deficits in the studied family, an inspection ofTable 1 (columns 5–12) necessitates a more detailed analysis.

Of the 28 tested subjects, four individuals displayed pronounceddeficits on at least one subtest of the object part of the VOSP, indicatedby a z score below �2. Their scores suggest difficulties with basic-levelobject identification. The most pronounced deficits were seen on thesubtest Object Decision, with a z score as low as �3.5 for one subject(F48). On this subtest, subjects were asked to select the silhouette of areal object out of an array that included three distractor items(nonsense objects). At the group level, the means in the subtestsObject Decision and (to a lesser extent) Silhouettes both revealed

performances significantly below average (see Table 1). Thus, as withface processing, the family as a whole displayed a general impairmentin basic-level object identification skills. Regarding spatial perception(subtests Number Location and Cube Analysis; Table 1, column 11 and12), three subjects displayed deficits on the subtest Number Location.However, the group means for the two subtests of the space percep-tion part of the VOSP were either average (Number Location) or evenabove average (Cube Analysis).

Although the group means for the four RCFT measures were notsignificantly impaired, the profiles of M28, F23, M21, M50, F19b, andagain F48 revealed major deficits in perceptual organization skills andfigural memory, indicated by z scores below �2 on at least onemeasure of the RCFT (Table 1). Whereas in some of these profiles, poorresults appeared in isolation (both, F19b and M50 scored well withinnormal ranges on three out of the four RCFT measures), three othersubjects displayed consistently abnormal results on the RCFT (F48, F23,M28), although they were unimpaired in face recognition as measuredby the CFMT. A clinical interpretation of these values would suggest amoderately-to-severely impaired memory for complex visual stimuli(Meyers & Meyers, 1995). Fig. 2 exemplary shows the results of subjectF48 on three measures of the RCFT and gives an impression of howitems seem to be recalled in an unsuccessful, feature-based mannerthat misses the overall organization of the figure. z Scores for testingsignificance of test differences moreover indicate that F48 shows asignificant discrepancy between face recognition as measured by theCFMTand nonface tasks. Also for F23 andM28, the difference betweenCFMT and RCFT immediate recall score is highly unlikely to beexplainable by measurement error (Table 2).

3.5. Correlations between general visual processing and faceperception skills

Given both, the accumulation of these diverse visuoperceptualimpairments in the investigated sample, as well as our finding ofsubjects with impaired scores in the CFPT but not in the CFMT, wefurther explored possible relationships between the general visualprocessing tasks and the face processing tasks by means of acorrelational analysis.

Whereas the CFMT did not share a significant amount of variancewith any nonface task in our battery, the CFPT upright was signifi-cantly correlated with the immediate recall trial of the RCFT, r(26)¼� .41, p¼ .031, suggesting a relationship between nonface per-ceptual organization skills and the process of judging perceptualsimilarities in faces: Subjects with high recall scores on the RCFTgenerally made fewer errors on the CFPT, and vice versa. Correlationsbetween all three RCFT drawing scores and the inverted face trials ofthe CFPT (CFPT inverted) were even stronger copy trial, r(26)¼� .63,po.001; immediate recall, r(26)¼� .55, p¼ .003; delayed recall, r(26)¼� .50, p¼ .006. This further supports the interpretation of aconnection between general perceptual organization skills and per-formance in the CFPT, independent of whether the faces that arecompared in the CFPT are upright or inverted. No other correlationsbetween face- and nonface tests reached significance.

4. Discussion

Previous family studies on CP have reported mixed andeven conflicting results regarding associated cognitive deficitsapart from face recognition, thus concluding that CP may be aheterogeneous clinical condition (Schmalzl et al., 2008; Lee et al.,2010). However it is unclear whether differences in the employeddiagnostic criteria for CP may account for some of the previouslyreported heterogeneity. Consensual clinical diagnostic criteria forCP based on standardized neuropsychological tests and extensivenormative data have been proposed to improve both, diagnostic

Table 3Intercorrelations of face processing measures.

Measure CFMT total CFPT uprighterrors

CFPTinversion effect

CFPT upright errors � .57nn

CFPT inversion effect .30 � .63nn

Self reported deficits � .50nn .20 � .05

Note. Results of correlational analyses with data from all 28 subjects. Note that forthe CFPT upright and the self report, higher scores indicate greater impairment.npo .05. nnpo .01.

A. Johnen et al. / Neuropsychologia 58 (2014) 52–6358

accuracy for single-case classification of CP as well as compar-ability between studies (Bowles et al., 2009). Here, following thesediagnostic guidelines, we present neuropsychological data of asample of 28 high functioning subjects from one family in whichmany members reported difficulties with face recognition. Weadministered standardized tests for face recognition (CFMT) andface perception (CFPT) and analyzed patterns of impairmentswithin the family. Additionally, two tests that assess more basicvisual abilities and perceptual organization skills (VOSP and RCFT)were employed in order to exclude the possibility of general visualintegrity impairments in suspected CPs. We also administered aself-report questionnaire on face recognition deficits in order todirectly compare objective and subjective data on face processingabilities.

4.1. Patterns of impairments and heterogeneity of face processingdeficits within the family

On an individual basis, we classified three persons – a fatherand his two sons (M46, M16b, M13) – as CP with severe impair-ments in both, face recognition and facial similarity judgmentabilities. Their poor performances on the CFMT and the reduced orabsent perceptual face inversion effect (CFPT) suggest a deficientholistic processing of faces in these subjects. Importantly, thesethree subjects showed no deficits in basic-level object identifica-tion (VOSP) or general perceptual organization of a configuralnonface stimulus (RCFT). Although higher-order object processingskills (e.g., within-category discrimination) were not tested here,this result is compatible with work that views CP as a higher-order

visual processing deficit for faces (Duchaine & Nakayama, 2005,2006a; Duchaine, Yovel, Butterworth, & Nakayama, 2006;Lobmaier, Bol̈te, Mast, & Dobel, 2010). Moreover, the highly similartest profiles in these first-degree relatives strongly suggest afamilial transmission of face processing skills, which is in linewith recent studies both in the general population (Wilmer et al.,2010; Zhu et al., 2010) and in families with prosopagnosicmembers (e.g., Kennerknecht et al., 2008b; Duchaine et al.,2007a; Lee et al., 2010).

In addition to this nonambiguous triplet of CPs, four morefamily members (M23, M16a, M27, M21) also displayed majorimpairments in both, face perception and face recognition, how-ever did not reach conservative cut-off values for CP and/or met anexclusion criterion for such a classification (e.g., major clinicalimpairment in a nonface visual memory task). Another group offamily members displayed difficulties only in judging facial simi-larity as measured by the CFPT but not in face recognition, the corefeature of prosopagnosia.

In summary, our study revealed evidence for a severe type offace processing deficit in spite of normal performance in severalnonface tasks of visual integrity (i.e., CP). However, we also foundevidence for less severe and more heterogeneous dysfunctions, inwhich specific processes of face processing (i.e., face recognition orfacial similarity judgment) can be differentially impaired andwhich can additionally be accompanied by deficits in generalvisuoperceptual processing. This latter finding relates to severalstudies that have concluded that CP itself may be a heterogeneouscondition and that different subtypes can manifest themselves atdifferent stages in the process of face cognition (Le Grand et al.,

Fig. 2. Subject F48's performance on the RCFT trials.

A. Johnen et al. / Neuropsychologia 58 (2014) 52–63 59

2006; Schmalzl et al., 2008; Stollhoff et al., 2011). In an influentialfamily study, Schmalzl et al. (2008) reported highly variablecognitive profiles in members of one family. They were able toshow that the inability to recognize family members assessed witha tailor-made experimental test was not consistently related tospecific deficits in face processing tasks across all subjects. Otherfamily studies also reported on cases with diverse visuoperceptualdeficits regarding associated cognitive functions such as within-class object discrimination or nonface global-local tasks (Duchaineet al., 2007b; Lee et al., 2010).

Given these results, heterogeneous neuropsychological profileseven within one family seem to be the norm rather than anexception, including diverse deficits in standard tasks of generalvisual processing. This result emphasizes the necessity of diag-nostic criteria for CP that include the assessment of general visualprocessing skills in order to improve single-case diagnostic accu-racy and exclude ambiguous cases that may be better described byother impairments.

4.2. Familial contributions to face processing abilities

Even apart from the three diagnosed CPs from one core family,we found strong evidence for a familial contribution to faceprocessing skills. For the 21 family members who were notdiagnosed as CP and were not even suggestive for this diagnosis(i.e., were not even classified as bad recognizers), the group meanwas nevertheless significantly below average on the two faceprocessing tasks. This finding has important implications regard-ing the distinctiveness of CP as a clinical condition: Althoughrelatives of CPs may not be impaired in a clinical sense (i.e., do notreach single-case cut-off scores in standardized tests), they seemto be prone to subtle impairments of face processing that manifestthemselves on sensitive behavioral tasks such as the CFMT or theCFPT. As such, the presented family sample provides preliminaryevidence for a generally reduced group performance on faceprocessing tasks in close relatives of CPs. In line with Wilhelmet al. (2010), our results thus argue for a continuum of faceprocessing skills rather than distinct categories. CPs in this sensesimply reflect the low end of this continuum, and because faceprocessing skills have a substantial genetic base, these extremedeviations from the norm are more likely to be found withincertain families. Estimates of CP prevalence at about 2.5%(Kennerknecht et al., 2006, 2008a; Bowles et al., 2009) alsosupport this view of a normal distribution of face processing skills,given that this portion approximately corresponds to the lowerend of a normal distribution, i.e. 2 SD below the mean. Reports onpeople who achieve extraordinarily high scores on the CFMT andCFPT provide further support for the continuum hypothesis(Russell et al., 2009).

4.3. Gender differences in face recognition abilities

Due to the fact all three diagnosed CP and all of the badrecognizers in our sample were males, we tested the genderdistribution of face processing deficits. A gender effect emergedbut only if the two groups (CP and bad recognizers) were merged,suggesting that male subjects were more likely to be classified asCP or bad recognizers in our sample. This result is in line withconverging evidence suggesting, that females may show a slightadvantage in recognizing faces over males and that this effect is atleast partly independent of the stimulus-gender (Bowles et al.,2009; Rehnmann & Herlitz, 2007). Moreover, it has been suggestedthat acquired prosopagnosia may be more frequent in men due todifferences in cerebral organization (Mazzuchi & Biber, 1983). Incontrast to these findings, Kennerknecht et al. (2008b) found thatmore women than men were impaired in a large survey of 38

families with face recognition impairments. Due to major differ-ences in the applied diagnostic criteria for CP, these results arehowever, hard to compare and the current sample size is too smallto draw further conclusions from our results. In summary, theaspect of gender distribution among CP can only be addressed byan epidemiological study in a much larger population.

4.4. Self-reports as a screening instrument for CP

In order to evaluate the validity of self-reports as a possiblescreening procedure for CP, we directly compared scores on astandardized self-report questionnaire on face recognition withneuropsychological tests. All three classified CPs in the family alsorated their own face processing skills as majorly impaired. More-over, we found a moderate correlation of r¼ .50 between the CFMTand the self-report questionnaire in the current sample. Indyslexia, a developmental disorder of similar cognitive specificity,correlation sizes between neuropsychological literacy measuresand self-report data are comparable in size (e.g., Snowling et al.,2012). We further calculated sensitivity and specificity of theemployed self-report questionnaire and found the use of self-report questionnaires as a screening instrument for CP was well-founded. Nevertheless, a diagnosis of CP solely based on the shortscreening questionnaire employed here would have led to severalfalse-positive classifications (e.g., M52, M28, M49). This resultstrongly argues for an in-depth neuropsychological investigationto validate self-ascribed face processing deficits.

4.5. General visuoperceptual deficits

To our surprise, the evaluation of the nonface general visualprocessing tasks revealed visuoperceptual deficits in many mem-bers of the family. Whereas some subjects displayed symptoms ofa mild object agnosia, others showed a pattern of impaired figuralmemory and perceptual organization skills, and some had majordifficulties with both nonface visual tasks. These results are evenmore striking, considering that the employed general visualprocessing tasks were meant to serve as rough screening measuresfor neuropsychological dysfunctions of general visual integrity. Aswith the face processing tasks, the group means on two of thesemeasures for basic-level object identification (VOSP Object Deci-sion, VOSP Silhouettes) were significantly below average, againsuggesting a familial aggregation of impairments on these tasks.

Comparing these results to other studies, some cases of CP inthe literature have shown deficits in object processing skills(Behrmann et al., 2005; Duchaine et al., 2007a; Lee et al., 2010).However, these deficits were limited to poor performance onhigher-order within-category discrimination tasks (houses orcars). Relatedly, Germine, Cashdollar, Düzel, and Duchaine (2011)recently reported on a developmental case of impaired within-class object recognition but normal face recognition abilities.However, subject AW was not impaired in either figural memoryor perceptual organization skills (RCFT), or in basic-level objectidentification assessed by tasks comparable to the VOSP. Bycontrast, our study revealed major deficits in close relatives ofCPs in an object identification task (VOSP) and/or in the efficientorganization of a configural stimulus (RCFT). Such deficits werealso found in family members who showed only limited evidencefor CP, instead of the full spectrum (i.e., impaired facial similarityjudgment but normal face recognition) emphasizing the need totest these more basic nonface abilities in suspected CPs.

Considering that there was no indication of a neurological orpsychiatric disorder that could explain these results for any of thefamily members, the cause of their impaired figural memory andperceptual organization skills remains unclear. Although theobservation of a high prevalence of both deficient general

A. Johnen et al. / Neuropsychologia 58 (2014) 52–6360

visuoperceptual processing and impaired face processing withinone family may be just a coincidence, these results neverthelesssuggest a mutual risk factor in this specific family for visuoper-ceptual deficits in general, one of which may be CP. Whether theparallel occurrence of face recognition impairments and generalvisuoperceptual deficits is idiosyncratic to this sample, or can befound in other families with prosopagnosic members has to beexplored in future studies.

4.6. Differences between CFMT and CFPT and their associations withtasks of general visual processing

Interestingly, analyses of individual differences showed a signifi-cant discrepancy between CFMT and CFPT in two family members(F48, M18). Moreover, a group of subjects with impairments onnonface tasks performed normally in face recognition (CFMT) butfailed on the CFPT (F48, M50, F23, M28). These results may indicatethat judging facial similarities in the CFPT is stronger affected bygeneral visuoperceptual deficits than face recognition in the CFMTeven though both tasks make use of the same stimuli set. Thispreliminary interpretation was supported by a significant correlationbetween the CFPT and the RCFT at the group-level, which importantlydid not emerge for the CFMT. Correlations between RCFT and CFPTwere even stronger for CFPT inverted trials, suggesting that thehypothesized influence of general visuoperceptual abilities on theCFPT is independent of whether the stimulus that is judged is a face oran inverted face. Moreover, the CFMT and CFPT were strongly but notperfectly correlated, giving rise to potentially modulating influences onone task but not the other. Thus, our results suggest that the CFPT maybe considered a task that is at least not exclusively sensitive to face-specific perceptual deficits, but also to more general visuoperceptualdeficits for objects with a complex featural structure.

One possible explanation for the effects of general visuoper-ceptual deficits on the CFPT but not on the CFMT regards theinfluence of featural elements in the two tasks. Although stimuli ofthe CFMT and CFPT have been designed to avoid comparisons offacial features, a task like the CFPT, in which a target-face ispresent throughout the trial, may trigger attempts of featurematching, as previously shown for a simple face matching task(Rotshtein, Geng, Driver, & Dolan, 2007). A further potentialexplanation regards the influence of timed responses. Whereasthe CFMT is untimed, the CFPT has a time limit of 1 min. per trial.Since speed is also considered as a component of face cognitionthat is widely independent of face recognition and face perceptionabilities (Wilhelm et al., 2010), it may be that individual differ-ences in face processing speed account for the found discrepanciesbetween CFMT and CFPT. Another methodological difference is thesize of the pictures as they were presented on the screen. Whilethe CFMT displays three faces each trial, the CFPT displays six,leading to smaller picture sizes which might have adverselyaffected identity sorting.3

4.7. Limitations

With the current study, we aimed at contributing to the body ofevidence on familial aggregation of face processing deficits byemploying suggested clinical diagnostic guidelines for CP in a largefamily with subjective face recognition problems. In particular, weassessed the level of heterogeneity regarding patterns of impair-ments on neuropsychological tests for face processing and moregeneral visuoperceptual abilities. However, several limitationsneed to be considered when interpreting our results.

First, although an in-depth semistructured interview revealedno indication for face processing problems among the spouses ofthe first generation of the family, the lack of a formal neuropsy-chological testing of this group certainly limits an interpretation ofthe tested parents' genetic contributions to face processing skillsof their children. Objective neuropsychological data in completefamily pedigrees from multiple generations are needed in order tobetter understand the genetic contribution to the development ofCP.

Second, although basic nonface visuoperceptual integrity andnonface visual memory was tested and found to be intact in alldiagnosed CP, we have no information on various other higher-order nonface visual processing abilities such as within-categorydiscrimination of objects (e.g., Dennett et al., 2012; McGugin,Richler, Herzmann, Speegle & Gauthier, 2012). This informationwould have been interesting in order to characterize otherpossible associated impairments in individuals classified as CP,since mixed results regarding such abilities have also beendescribed previously (e.g., Lee et al., 2010; Duchaine et al.,2007b). Importantly, our finding of impairments in basic nonfacevisuoperceptual functions in many family members is unaffectedby the lack of information on within-category discrimination. Thisfinding underlines the importance of conservative classificationsof CP and the need to control for basic nonface visuoperceptualintegrity in suspected subjects. Future studies are needed toelucidate nonface visual abilities at higher processing levels inindividuals that fulfill current criteria for CP.

A third limitation regards the reported correlations and pro-posed functional association between face and nonface tasksemployed in this study. These correlations are based on a rathersmall number of subjects and thus care should be taken whentransferring our conclusions to other samples. Although all mea-sures were standardized, the investigated family may also beexceptional with regard to correlations between the tested neu-ropsychological domains. Thus, the degree to which these pro-posed associations and dissociations can be generalized to othersamples needs to be investigated on a larger scale in futurestudies.

4.8. Conclusions

(a) We employed established face processing tests as well asstandard nonface tests of general visual integrity in an extendedfamily sample suggestive of CP. Despite highly homogeneousgenetic and environmental conditions, only three first-degreerelatives fulfilled conservative clinical criteria for CP with severeimpairments in face processing, a reduced or absent face inversioneffect, and at least grossly intact nonface visuoperceptual proces-sing. This result highly emphasizes the importance of clinicaldiagnostic criteria and exclusion based on tests for general visualintegrity. (b) Nevertheless, most of the remaining family members(i.e., subjects that were not indicative of CP based on publishedcriteria) also showed considerable deficits in face processingabilities, and the family's mean scores in two standardized faceprocessing tests were significantly lower than the average of thegeneral population. This suggests that face processing abilities areinfluenced by a shared familial factor and are represented in acontinuous (possibly normal) distribution. (c) Self-report datafrom a standardized questionnaire were moderately correlatedwith face recognition scores and showed a high sensitivity andacceptable specificity regarding diagnoses with CP. (d) Addition-ally, in some family members with normal face recognition, wenonetheless found severe impairments in perceptual organizationof nonface stimuli and figural memory deficits as well as asignificantly reduced group mean in an object identification task.These results suggest the existence of a mutual genetic and/or3 We thank an anonymous reviewer for providing this possible explanation.

A. Johnen et al. / Neuropsychologia 58 (2014) 52–63 61

environmental risk factor for visuoperceptual impairments in thisfamily, which may manifest itself either in specific and severedifficulties with recognizing familiar faces (i.e., CP) or in moregeneral visuoperceptual deficits. (e) Both, on a single-case level aswell as in a correlational analysis, these general aspects ofvisuoperceptual processing were statistically independent of facerecognition skills as measured by the CFMT. In contrast, the CFPTwas correlated with tasks of nonface visuoperceptual abilites. Poorresults on the CFPT may therefore not be specific for CP. Futureresearch should concentrate on extensive assessment of basicperceptual organization skills in persons with face processingdeficits and also in their close family members in order to replicateour findings and to disentangle the genetic and developmentalfactors that underly face- and nonface visuoperceptual deficits.

References

Akshoomoff, N. A., & Stiles, J. (1995). Developmental trends in visuospatial analysisand planning: I. Copying a complex figure. Neuropsychology, 9(3), 364–377.

Anderson, P., Anderson, V., & Garth, J. (2001). Assessment and development oforganizational ability: The Rey Complex Figure Organizational Strategy Score(RCF-OSS). The Clinical Neuropsychologist, 15(1), 81–94.

Avidan, G., Tanzer, M., & Behrmann, M. (2011). Impaired holistic processing incongenital prosopagnosia. Neuropsychologia, 49(9), 2541–2552.

Behrmann, M, & Avidan, G. (2005). Congenital prosopagnosia: Face-blind frombirth. Trends in Cognitive Sciences, 2(1), 245–268.

Behrmann, M., Avidan, G., Marotta, J. J., & Kimchi, R. (2005). Detailed exploration offace-related processing in congenital prosopagnosia: 1. Behavioral findings.Journal of Cognitive Neuroscience, 17(7), 1130–1149.

Bonello, P. J., Rapport, L. J., & Millis, S. R. (1997). Psychometric properties of thevisual object and space perception battery in normal older adults. The ClinicalNeuropsychologist, 11(4), 436–442.

Bowles, D. C., McKone, E., Dawel, A., Duchaine, B., Palermo, R., Schmalzl, L., et al.(2009). Diagnosing prosopagnosia: Effects of ageing, sex, and participant–stimulus ethnic match on the Cambridge Face Memory Test and CambridgeFace Perception Test. Cognitive Neuropsychology, 26(5), 423–455.

Busigny, T., & Rossion, B. (2011). Holistic processing impairment can be restricted tofaces in acquired prosopagnosia: Evidence from the global/local Navon effect.Journal of Neuropsychology, 5(1), 1–14.

De Haan, E. H. F. (1999). A familiar factor in the development of face recognitiondeficits. Journal of Clinical and Experimental Neuropsychology, 21(3), 312–315.

Deckersbach, T., Savage, C. R., Henin, A., Mataix-Cols, D., Otto, M. W., Wilhelm, S.,et al. (2000). Reliability and validity of a scoring system for measuringorganizational approach in the Complex Figure Test. Journal of Clinical andExperimental Neuropsychology, 22(5), 640–648.

Dennett, H. W., McKone, E., Tvashmi, R., Hall, A., Pidcock, M., Edwards, M., et al.(2012). The Cambridge Car Memory Test: A task matched in format to theCambridge Face Memory Test, with norms, reliability, sex differences, dissocia-tions from face memory, and expertise effects. Behavioral Research Methods, 44,587–605.

Dinkelacker, V., Grüter, M., Klaver, P., Grüter, T., Specht, K., Weis, S., et al. (2010).Congenital prosopagnosia: Multistage anatomical and functional deficits in faceprocessing circuitry. Journal of Neurology, 258(5), 770–782.

Dobel, C., Bol̈te, J., Aicher, M., & Schweinberger, S. R. (2007). Prosopagnosia withoutapparent cause: Overview and diagnosis of six cases. Cortex, 43(6), 718–733.

Dobel, C., Putsche, C., Zwitserlood, P., & Junghof̈er, M. (2008). Early left-hemisphericdysfunction of face processing in congenital prosopagnosia: An MEG study.PLoS One, 3(6), e2326.

Duchaine, B., Germine, L., & Nakayama, K. (2007a). Family resemblance: Ten familymembers with prosopagnosia and within-class object agnosia. Cognitive Neu-ropsychology, 24(4), 419–430.

Duchaine, B., & Nakayama, K. (2004). Developmental prosopagnosia and the BentonFacial Recognition test. Neurology, 62(7), 1219–1220.

Duchaine, B., & Nakayama, K. (2005). Dissociations of face and object recognition indevelopmental prosopagnosia. Journal of Cognitive Neuroscience, 17(2),249–261.

Duchaine, B., & Nakayama, K. (2006a). Developmental prosopagnosia: A window tocontent-specific face processing. Current Opinion in Neurobiology, 16(2),166–173.

Duchaine, B., & Nakayama, K. (2006b). The Cambridge Face Memory Test: Resultsfor neurologically intact individuals and an investigation of its validity usinginverted face stimuli and prosopagnosic participants. Neuropsychologia, 44(4),576–585.

Duchaine, B., & Weidenfeld, A. (2003). An evaluation of two commonly used tests ofunfamiliar face recognition. Neuropsychologia, 41(6), 713–720.

Duchaine, B., Yovel, G., Butterworth, E. J., & Nakayama, K. (2006). Prosopagnosia asan impairment to face-specific mechanisms: Elimination of the alternativehypotheses in a developmental case. Cognitive Neuropsychology, 23(5), 714–747.

Duchaine, B., Yovel, G., & Nakayama, K. (2007b). No global processing deficit in theNavon task in 14 developmental prosopagnosics. Social Cognitive and AffectiveNeuroscience, 2(2), 104–113.

Farah, M. J. (1996). Is face recognition “special”? Evidence from neuropsychology.Behavioural Brain Research, 76(1–2), 181–189.

Farah, M. J., Wilson, K. D., Drain, H. M., & Tanaka, J. N. (1995). The inverted faceinversion effect in prosopagnosia: Evidence for mandatory, face-specific per-ceptual mechanisms. Vision Research, 35(14), 2089–2093.

Farah, M. J., Wilson, K. D., Drain, H. M., & Tanaka, J. N. (1998). What is “special”about face perception? Psychological Review, 105(3), 482–498.

Gainotti, G. (2010). Not all patients labeled as “prosopagnosia” have a realprosopagnosia. Journal of Clinical and Experimental Neuropsychology, 32(7),763–766.

Galaburda, A. M., & Duchaine, B. (2003). Developmental disorders of vision.Neurologic Clinics, 21(3), 687–707.

Germine, L., Cashdollar, N., Düzel, E., & Duchaine, B. (2011). A new selectivedevelopmental deficit: Impaired object recognition with normal face recogni-tion. Cortex, 47(5), 598–607.

Grueter, M., Grueter, T., Bell, V., Horst, J., Laskowski, W., Sperling, K., et al. (2007).Hereditray prosopagnosia: The first case series. Cortex, 43(6), 734–749.

Haxby, J., Hoffman, E., & Gobbini, M. (2000). The distributed human neural systemfor face perception. Trends in Cognitive Sciences, 4(6), 223–233.

Herzmann, G., & Danthiir, V. (2008). Toward a comprehensive test battery for facecognition: Assessment of the tasks. Behavior, 40(3), 840–857.

Jones, R., & Tranel, D. (2001). Severe developmental prosopagnosia in a child withsuperior intellect. Journal of Clinical and Experimental Neuropsychology, 23(3),265–273.

Kanwisher, N., McDermott, J., & Chun, M. M. (1997). The fusiform face area: Amodule in human extrastriate cortex specialized for face perception. The Journalof Neuroscience, 17(11), 4302–4311.

Kennerknecht, I., Grüter, T., Welling, B., & Wentzek, S. (2006). First report ofprevalence of non-syndromic hereditary prosopagnosia (HPA). American Journalof Medical Genetics, 1622, 1617–1622.

Kennerknecht, I., Ho, N. Y., & Wong, V. C. N. (2008a). Prevalence of hereditaryprosopagnosia (HPA) in Hong Kong Chinese population. American Journal ofMedical Genetics, 146A(22), 2863–2870.

Kennerknecht, I., Kischka, C., Stemper, C., Elze, T., & Stollhoff, R. (2011). Heritabilityof face recognition. In: T. Barbu (Ed.), Face analysis, modeling and recognitionsystems (163–188). Rijeka, Croatia: InTech-Open Access Publisher.

Kennerknecht, I., Pluempe, N., Edwards, S., & Raman, R. (2007). Hereditaryprosopagnosia (HPA): The first report outside the Caucasian population. Journalof Human Genetics, 52(3), 230–236.

Kennerknecht, I., Pluempe, N., & Welling, B. (2008b). Congenital prosopagnosia—Acommon hereditary cognitive dysfunction in humans. Frontiers in Bioscience, 13,3150–3158.

Kramer, J. H., & Wells, A. M. (2004). The role of perceptual bias in complex figurerecall. Journal of Clinical and Experimental Neuropsychology, 26(6), 838–845.

Kress, T., & Daum, I. (2003). Developmental prosopagnosia: A review. BehaviouralNeurology, 14(3), 109–121.

Lange, J., De Lussanet, M., Kuhlmann, S., Zimmermann, A., Lappe, M., Zwitserlood, P.,et al. (2009). Impairments of biological motion perception in congenitalprosopagnosia. PLoS One, 4(10), e7414.

Le Grand, R., Cooper, P. A., Mondloch, C. J., Lewis, T. L., Sagiv, N., De Gelder, B., et al.(2006). What aspects of face processing are impaired in developmentalprosopagnosia? Brain and Cognition, 61(2), 139–158.

Le Grand, R., Mondloch, C. J., Maurer, D., & Brent, H. P. (2004). Impairment inholistic face processing following early visual deprivation. Psychological Science,15(11), 762–768.

Lee, Y., Duchaine, B., Wilson, H. R., & Nakayama, K. (2010). Three cases ofdevelopmental prosopagnosia from one family: Detailed neuropsychologicaland psychophysical investigation of face processing. Cortex, 46(8), 949–964.

Lezak, M. D., Howieson, D. B., Bigler, E. D., & Tranel, D. (2012). Neuropsychologicalassessment (5th ed.). New York: Oxford University Press.

Lobmaier, J. S., Bol̈te, J., Mast, F. W., & Dobel, C. (2010). Configural and featuralprocessing in humans with congenital prosopagnosia. Advances in CognitivePsychology, 6, 23–34.

Maurer, D., Grand, R. L., & Mondloch, C. J. (2002). The many faces of configuralprocessing. Trends in Cognitive Sciences, 6(6), 255–260.

Mazzuchi, A., & Biber, C. (1983). Is prosopagnosia more frequent in males than infemales? Cortex, 19(4), 509–516.

McGugin, R. W., Richler, J. J., Herzmann, G., Speegle, M., & Gauthier, I. (2012). TheVanderbilt Expertise Test reveals domain-general and domain-specific sexeffects in object recognition. Vision Research, 69, 10–22.

Meyers, J., & Meyers, K. (1995). Rey Complex Figure and Recognition Trial Professionalmanual. Odessa FL: Psychological Assessment Resources.

Moscovitch, M., Winocur, G., & Behrmann, M. (1997). What is special about facerecognition? Nineteen experiments on a person with visual object agnosia anddyslexia but normal face recognition. Journal of Cognitive Neuroscience, 9,555–604.

Newman, P. D., & Krikorian, R. (2001). Encoding and complex figure recall. Journal ofthe International Neuropsychological Society, 7(6), 728–733.

Omar, R., Rohrer, J. D., Hailstone, J. C., & Warren, J. D. (2011). Structural neuroa-natomy of face processing in frontotemporal lobar degeneration. Journal ofNeurology, Neurosurgery and Psychiatry, 82(12), 1341–1343.

A. Johnen et al. / Neuropsychologia 58 (2014) 52–6362

Palermo, R., Willis, M. L., Rivolta, D., McKone, E., Wilson, C. E., & Calder, A. J. (2011).Impaired holistic coding of facial expression and facial identity in congenitalprosopagnosia. Neuropsychologia, 49(5), 1226–1235.

Payne, R. W., & Jones, H. G. (1957). Statisics for the investigation of individual cases.Journal of Clinical Psychology, 13(2), 115–121.

Ramlall, S., Chipps, J., Bhigjee, A. I., & Pillay, P. J. (2013). The sensitivity andspecificity of subjective memory complaints and the Subjective Memory RatingScale, Deterioration Cognitive Observee, Mini-Mental State Examination, Six-Item Screener and Clock Drawing Test in dementia screening. Dementia andGeriatric Cognitive Disorders, 36, 119–135.

Rehnmann, J., & Herlitz, A. (2007). Women remember more faces than men do. ActaPsychologica, 124, 344–355.

Rey, A. (1941). L'examen psychologique dans les cas d'encéphalopathie trauma-tique. Les problems./The psychological examination in cases of traumaticencephalopathy. Problems. Archives de Psychologie, 28, 215–285.

Robbins, R., & McKone, E. (2007). No face-like processing for objects-of-expertise inthree behavioural tasks. Cognition, 103(1), 34–79.

Rotshtein, P., Geng, J. J., Driver, J., & Dolan, R. J. (2007). Role of features and second-order spatial-relations in face discrimination, face recognition, and individualface skills: Behavioral and fMRI data. Journal of Cognitive Neuroscience, 19(9),1435–1452.

Russell, R., Duchaine, B., & Nakayama, K. (2009). Super-recognizers: People withextraordinary face recognition ability. Psychonomic Bulletin & Review, 16(2),252–257.

Schmalzl, L, Palermo, R., & Coltheart, M. (2008). Cognitive heterogeneity ingenetically based prosopagnosia: A family study. Journal of Neuropsychology,2(1), 99–117.

Schmalzl, L, Palermo, R., Harris, I. M., & Coltheart, M. (2009). Face inversionsuperiority in a case of prosopagnosia following congenital brain abnormal-ities: What can it tell us about the specificity and origin of face-processingmechanisms? Cognitive Neuropsychology, 26(3), 286–306.

Snowling, M., Dawes, P., Nash, H., & Hulme, C. (2012). Validity of a protocol for adultself-report of dyslexia and related difficulties. Dyslexia, 18(1), 1–15.

Stollhoff, R., Jost, J., Elze, T., & Kennerknecht, I. (2011). Deficits in long-termrecognition memory reveal dissociated subtypes in congenital prosopagnosia.PLoS One, 6(1), e15702.

Tanaka, J. W., & Farah, M. J. (1993). Parts and wholes in face recognition. QuarterlyJournal of Experimental Psychology. A, Human Experimental Psychology, 46(2),225–245.

Tree, R. R., & Wilkie, J. (2010). Face and object imagery in congenital prosopagnosia:A case series. Cortex, 46(9), 1189–1198.

Van Belle, G., De Graef, P., Verfaillie, K., Busigny, T., & Rossion, B. (2010). Whole nothole: Expert face recognition requires holistic perception. Neuropsychologia, 48(9), 2620–2629.

Von Cramon, D. Y., Mai, N., & Ziegler, W. (1995). Neuropsychologische Diagnostik.London: Chapman and Hall.

Warrington, E. K., & James, M. (1992). Testbatterie für visuelle Objekt- undRaumwahrnehmung. Bury St Edmunds: Thames Valley Test Company.

Weigelt, S., Koldewyn, K., & Kanwisher, N. (2012). Face identity recognition inautism spectrum disorders: A review of behavioral studies. Neuroscience andBiobehavioral Reviews, 36(3), 1060–1084.

Wilhelm, O., Herzmann, G., Kunina, O., Danthiir, V., Schacht, A., & Sommer, W.(2010). Individual differences in perceiving and recognizing faces—One elementof social cognition. Journal of Personality and Social Psychology, 99(3), 530–548.

Wilmer, J. B., Germine, L., Chabris, C. F., Chatterjee, G., Williams, M., Loken, E., et al.(2010). Human face recognition ability is specific and highly heritable. Proceed-ings of the National Academy of Sciences, 107(11), 5238–5241.

Wilmer, J. B., Germine, L., Chabris, C. F., Gerbasi, M., & Nakayama, K. (2012).Capturing specific abilities as a window into human individuality: The exampleof face recognition. Cognitive Neuropsychology, 29(5–6), 360–392.

Wilson, C., Freeman, P., Brock, J., Burton, A., & Palermo, R. (2010). Facial identityrecognition in the broader autism phenotype. PLoS One, 5(9), e12876.

Yin, R. K. (1969). Looking at upside-down faces. Journal of Experimental Psychology,81(1), 141–145.

Young, A. W., De Haan, E. H. F., & Bauer, R. M. (2008). Face perception: A very specialissue. Journal of Neuropsychology, 2(1), 1–14.

Young, A. W., Hellawell, D., & Hay, D. C. (1987). Configural information in faceperception. Perception, 16(6), 747–759.

Zhu, Q., Song, Y., Hu, S., Li, X., Tian, M., Zhen, Z., et al. (2010). Heritability of thespecific cognitive ability of face perception. Current Biology, 20(2), 137–142.

A. Johnen et al. / Neuropsychologia 58 (2014) 52–63 63