Crossmodal priming of unfamiliar faces supports early interactions between voices and faces in person perception

ABSTRACT

Although faces and voices are important sources of information for person recognition, it is unclear whether these cues interact at a late stage to act as complementary, unimodal sources for person perception or whether they are integrated early on to provide a multisensory representation of a person in memory. Here we used a crossmodal associative priming paradigm to test whether unfamiliar voices which were recently paired with unfamiliar faces could subsequently prime familiarity decisions to the related faces. Based on our previous study, we also predicted that distinctive voices would enhance the recognition of faces relative to typical voices. In Experiment 1 we found that voice primes facilitated the recognition of related target faces at test relative to learned but unrelated voice primes. Furthermore, face recognition was enhanced by the distinctiveness of the paired voice primes. In contrast, we found no evidence of priming with arbitrary sounds (Experiment 2), confirming the special status of the pairing between voices and faces for person identification. In Experiment 3, we established that voice primes relative to no prime facilitated familiarity decisions to related faces. Our results suggest a strong association between newly learned voices and faces in memory. Furthermore, the distinctiveness effect found for voice primes on face recognition suggests that the quality of the voice can affect memory for faces. Our findings are discussed with regard to existing models of person perception and argue for interactions between voices and faces that converge early in a multisensory representation of persons in long-term memory.

KEYWORDS:

• Faces
• voices
• distinctiveness
• priming
• crossmodal

Acknowledgments

We thank Daniel Berger for help creating the auditory stimuli, and Franziska Hausmann, Karin Bierig and Saskia Kühnhold for help in programming the experiments and testing participants. We are grateful to our colleagues for allowing us to record their voices for use as stimuli.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

1 Participants failed to provide a response within this time limit in less than 0.5% of all trials.

2 For completeness, Pearson correlations were calculated for unrelated trials in Experiment 1 (there were no unrelated trials to analyse in Experiment 3). Both correlations, calculated for response times and accuracy data, were non-significant.

Additional information

Funding

This research was supported by the Max-Planck-Gesellschaft and by funding from Science Foundation Ireland [Grant No. 10/IN.1/I3003] awarded to FNN.