Normal composite face effects in developmental prosopagnosia

Abstract

Upright face perception is thought to involve holistic processing, whereby local features are integrated into a unified whole. Consistent with this view, the top half of one face appears to fuse perceptually with the bottom half of another, when aligned spatially and presented upright. This 'composite face effect' reveals a tendency to integrate information from disparate regions when faces are presented canonically. In recent years, the relationship between susceptibility to the composite effect and face recognition ability has received extensive attention both in participants with normal face recognition and participants with developmental prosopagnosia. Previous results suggest that individuals with developmental prosopagnosia may show reduced susceptibility to the effect suggestive of diminished holistic face processing. Here we describe two studies that examine whether developmental prosopagnosia is associated with reduced composite face effects. Despite using independent samples of developmental prosopagnosics and different composite procedures, we find no evidence for reduced composite face effects. The experiments yielded similar results; highly significant composite effects in both prosopagnosic groups that were similar in magnitude to the effects found in participants with normal face processing. The composite face effects exhibited by both samples and the controls were greatly diminished when stimulus arrangements were inverted. Our finding that the whole-face binding process indexed by the composite effect is intact in developmental prosopagnosia indicates that other factors are responsible for developmental prosopagnosia. These results are also inconsistent with suggestions that susceptibility to the composite face effect and face recognition ability are tightly linked. While the holistic process revealed by the composite face effect may be necessary for typical face perception, it is not sufficient; individual differences in face recognition ability likely reflect variability in multiple sequential processes.

Holistic face perception is impaired in developmental prosopagnosia

“…What do the current findings imply for the cognitive and neural mechanisms that are responsible for impaired face recognition in DP, and in particular for the role of holistic processing? Previous behavioural studies using different face matching paradigms (composite face and part-whole tasks) have shown that individuals with DP can have problems with holistic face processing, but that this is not always the case (Avidan et al, 2011;De Gutis et al, 2012;Biotti et al, 2017). Demonstrations that DPs may have normal holistic face processing abilities have contributed to the debate about the extent to which holistic face perception is related to face recognition ability in the general population (Richler, Cheung, & Gauthier, 2011;DeGutis, Wilmer, Mercado, & Cohan, 2013).…”

“…Studies with DP using this task have produced inconsistent results, with some reporting a reduced CFI for DPs (e.g. Avidan et al 2011), while others find no difference in the size of the CFI between DPs and control participants (e.g., Biotti et al, 2017). Clearer evidence for deficits in holistic face processing deficits for DPs comes from part-whole face matching tasks.…”

Normative data for two challenging tests of face matching under ecological conditions

“…Under simultaneous matching conditions observers are more likely to display speed-accuracy trade-offs compared to, for example, recognition tasks. Even highly impaired prosopagnosic individuals can achieve accurate performance at the expense of vastly prolonged RTs, which is why it is common practice to take into account RTs in studies of both healthy and impaired individuals (e.g., Behrmann, Avidan, Marotta, & Kimchi, 2005;Biotti et al, 2017;Lao, Vizioli, & Caldara, 2013;Marotta et al, 2002;Tarr, 2003;White et al, 2017). Indeed, according to many, "the notion of speed-accuracy tradeoff (SAT) is pedestrian" (Heitz, 2014; see also Luce, 1986).…”

Normative data for two tests of face matching under ecological conditions

“…Contrariwise, here we considered both accuracy and RTs. Aiming to account for the "pedestrian notion" of speedaccuracy trade-offs (Heitz, 2014, p.1; see also Luce, 1986), which can be found in healthy as well abnormal populations (Marotta et al, 2002;Tarr, 2003;Behrmann, Avidan, Marotta, & Kimchi, 2005;Herzmann, Danthiir, Schacht, Sommer, & Wilhelm, 2008;Ramon & Rossion, 2010;Ramon et al, 2010;Biotti et al, 2017;White et al, 2017;Geskin & Behrmann, 2018) and may vary across age (Hildebrandt, Herzmann, Sommer & Wilhelm, 2010), analyses were performed on a composite scores encapsulating both measures. Thus, alongside providing normative data for the YBT and FICST, their respective value was assessed relative to tests that have been previously used as "standard screening" tests (Noyes et al, 2018;Bobak et al, 2016b) of individual face processing skill.…”

“…recognition tasksmore likely to display speed-accuracy trade-offs. Even highly impaired prosopagnosic individuals can achieve accurate performance at the expense of vastly prolonged RTs, which is why it is common practice to take into account RTs in studies of both healthy and impaired individuals (e.g., Marotta et al, 2002;Tarr, 2003;Behrmann et al, 2005;Biotti et al, 2017;Lao, Vizioli, & Caldara, 2013;White et al, 2017). Indeed, according to many, "the notion of speed-accuracy tradeoff (SAT) is pedestrian" (Heitz, 2014, p.1; see also Luce, 1986).…”