A two-way window on face recognition: Reply to Breen et al

Abstract

The difference between Breen and colleagues’ model1xModels of face recognition and delusional misidentification: a critical review. Breen, N. et al. Cognit. Neuropsychol. 2000; 17: 55–71Crossref | PubMedSee all References and the revision illustrated by Ellis and Lewis2xCapgras delusion: a window on face recognition. Ellis, H.D. and Lewis, M.B. Trends Cognit. Sci. 2001; 5: 149–156Abstract | Full Text | Full Text PDF | PubMed | Scopus (179)See all References comes down to what the doubled-headed arrow means in the Breen et al. model, and its consequences. We take the arrow to imply a mandatory and automatic information pathway, just like all the other links – that is, the kinds of links that have been implemented in the interactive activation (IAC) model of face recognition3xUnderstanding face recognition with an interactive activation model. Burton, A.M. et al. Br. J. Psychol. 1990; 81: 361–380Crossref | PubMed | Scopus (67)See all References. (If this is not the case then what is represented by this double-headed arrow might be equivalent to the additional components in the Ellis and Lewis model and so there is no difference between the models.) The proposed mandatory pathway between the Person Identity Nodes (PINs) and affective-response box would mean that activation in these would reinforce each other until a steady state is established. A logical consequence, therefore, is that the effects of lesions following bifurcation would be circumvented. A prosopagnosic patient (lesioned between FRUs and PINs) should be able to access person-specific information from a face via a route through the affective-response module. In reality, prosopagnosics cannot access person-specific information from a face and so, we suggest, this route cannot be present.For the Capgras delusion, Breen et al. place the locus for the evaluation of anomalous experiences at the PINs. Famous faces, therefore, will produce a lower activation on the relevant PIN for Capgras patients because the additional information via the affective-response system is missing. Theories of associative priming, however, suggest that there are many other ways of changing the level of activation of a PIN. Prior presentation of an unrelated person, for example, will lead to lower levels of activation3xUnderstanding face recognition with an interactive activation model. Burton, A.M. et al. Br. J. Psychol. 1990; 81: 361–380Crossref | PubMed | Scopus (67)See all References. If reduced activation at the level of the PIN is sufficient to produce an anomalous experience then such an experience should be elicited in associative-priming experiments. Participants in such experiments, however, neither report that the target faces look odd nor, indeed, that they are impostors. Thus, reduced activation in the PINs cannot be considered a sufficient anomalous experience to evoke delusions. In order for there to be an anomalous experience, we argue that it is necessary to have some comparator between person recognition and affective responses to familiar faces. This comparator is the box labelled ‘Integrative device’ in the Ellis and Lewis model (Ref. 2xCapgras delusion: a window on face recognition. Ellis, H.D. and Lewis, M.B. Trends Cognit. Sci. 2001; 5: 149–156Abstract | Full Text | Full Text PDF | PubMed | Scopus (179)See all References2, p. 155).The two-route model, based on the functional aspects of Bauer's account4xAutonomic recognition of names and faces: a neuropsychological application of the Guilty Knowledge Test. Bauer, R. Neuropsychologia. 1984; 22: 457–469Crossref | PubMed | Scopus (362)See all References, is generally accepted as a starting point for an account of Capgras delusions, in spite of its neuroanatomical shortcomings. There exist differences of opinion, however, in how degraded information becomes an anomalous experience. Whereas Breen and colleagues suggest that this occurs at the PINs, Ellis and Lewis suggest that, logically, it requires some further comparator mechanism. Obviously, resolving this interesting issue is an area for further research.