Italian normative data and validation of two neuropsychological tests of face recognition: Benton Facial Recognition Test and Cambridge Face Memory Test

Abstract

The Benton Facial Recognition Test (BFRT) and Cambridge Face Memory Test (CFMT) are two of the most common tests used to assess face discrimination and recognition abilities and to identify individuals with prosopagnosia. However, recent studies highlighted that participant-stimulus match ethnicity, as much as gender, has to be taken into account in interpreting results from these tests. Here, in order to obtain more appropriate normative data for an Italian sample, the CFMT and BFRT were administered to a large cohort of young adults. We found that scores from the BFRT are not affected by participants' gender and are only slightly affected by participant-stimulus ethnicity match, whereas both these factors seem to influence the scores of the CFMT. Moreover, the inclusion of a sample of individuals with suspected face recognition impairment allowed us to show that the use of more appropriate normative data can increase the BFRT efficacy in identifying individuals with face discrimination impairments; by contrast, the efficacy of the CFMT in classifying individuals with a face recognition deficit was confirmed. Finally, our data show that the lack of inversion effect (the difference between the total score of the upright and inverted versions of the CFMT) could be used as further index to assess congenital prosopagnosia. Overall, our results confirm the importance of having norms derived from controls with a similar experience of faces as the "potential" prosopagnosic individuals when assessing face recognition abilities.

Normative accuracy and response time data for the computerized Benton Facial Recognition Test (BFRT-c)

“…In our study, the split-half reliability of the BFRT-c in accuracy rates was moderate, as in previous reports (Albonico et al, 2017;Christensen, Riley, Heffernan, Love, & Santa Maria, 2002). However, we believe that this was merely due to the small number of items (eight in each subpart of the test, for three responses in each item), and to the high variance in difficulty of the different items of the test (see Fig.…”

“…We report average performance on a computerized version of the BFRT of 44.81 out of 54 (83%), a score that is remarkably close to that of the healthy participants in the original study reporting the test (45.3 or 83.9%; Benton & Van Allen, 1968), but lower than the score obtained in a recent report of Italian normative data (47.12; Albonico et al, 2017). Standard deviation (3.44; i.e., 6.37%) is smaller in our sample than reported originally (4.5; i.e., 8.33%) and this could be due to the more homogeneous population tested here in terms of age range as compared to the wide age range covered in the original study (16-65 years of age; Benton & Van Allen, 1968).…”

“…We found that females outperformed males on the BFRT-c, a finding that is generally in agreement with the literature. Several studies showed that females perform better than males on face-processing tasks (e.g., Bowles et al, 2009;Heisz, Pottruff, & Shore, 2013;McBain, Norton, & Chen, 2009;Megreya, Bindemann, & Havard, 2011;Rehnman & Herlitz, 2007;Sommer, Hildebrandt, Kunina-Habenicht, Schacht, & Wilhelm, 2013), although this female advantage was not found in all studies, in particular in the original report of the BFRT (Benton & Van Allen, 1968; see also Albonico et al, 2017). This advantage is also observed sometimes only in specific conditions (see Weirich, Hoffmann, Meißner, Heinz, & Bengner, 2011), and could be specific for female faces (the so-called Bfemale own-sex bias^; see, e.g., Lewin & Herlitz, 2002;Lovén, Herlitz, & Rehnman, 2011;McKelvie, Standing, St Jean, & Law, 1993;Megreya et al, 2011) as observed here for accuracy rates.…”

“…For this reason, Benton always stated that low performance on the BFRT was not diagnostic of (acquired) prosopagnosia (Benton, 1980). Nakayama (2004, 2006b) reached the same conclusion for people diagnosed with life-long difficulties in individual face recognition-that is, so-called developmental prosopagnosia (or prosopdysgnosia; Rossion, in press)-showing that 73% of their group of participants could perform at normal accuracy levels on the BFRT (see also Albonico et al, 2017).…”

Contrasting shared- and specific-mechanism accounts of developmental prosopagnosia: A new approach

“…Use of the New Zealandic sample would thus also have made the group-based analyses, to be presented below, impossible due to too few observations. The individual data for each of the 261 participants in the alternative reference sample are given in the accompanying file: 'Supplementary Data' Even though we consider it correct not to use the New Zealandic sample to evaluate the generalizability of the present results, it clearly highlights that the outcome of the present procedure depends heavily on the reference sample used as cut-off scores can differ from sample to sample (Albonico, Malaspina, & Daini, 2017;Bowles et al, 2009). While this is true of all approaches that uses reference samples to classify abnormality it is nevertheless an aspect that requires careful attention if the procedure we suggest here should be tried out for real.…”