Purchase this article with an account.
Isabel Gauthier, Olivia S.-C. Cheung; How holistic processing is affected by working memory load. Journal of Vision 2005;5(8):378. https://doi.org/10.1167/5.8.378.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Holistic processing (HP) for faces (and objects of expertise) can be measured in tasks where subjects selectively attend to and compare a part (e.g., the mouth) of two faces while trying to ignore other parts that may vary. HP is evidenced by a congruency effect between the correct responses on the target and irrelevant parts. Last year we argued that encoding is not the locus of this effect because changing the configuration at encoding did not reduce HP (Gauthier et al., VSS 2003). Working memory load (WML) can increase processing of irrelevant distractors when the relevant and irrelevant information are different objects (e.g., de Fockert, 2001; Yi et al., 2004). Here we ask if WML would reduce selective attention to a face part (increase HP), especially for a WML during the matching judgment. A WML of three items was added to a sequential matching task with face composites during the first (S1) or the second (S2) face. In the main task, always using faces, subjects matched the top parts of the two stimuli, trying to ignore the bottom parts. In the load-at-study condition, three items were shown, then the S1, the WM probe, and finally the S2. In the load-at test condition, the S1 was presented, followed by the three items, then by the S2 and finally the WM probe. A control condition used scrambled items instead of the WML stimuli and no response was required. Faces and watches were used for the WML in different versions. A face WML reduced HP in all conditions and the reduction was larger in the load-at-test than the load-at-study condition. This supports the idea that manipulations at test affect HP more than those at encoding. The watch WML did not impact face HP at either study or test, presumably because watches did not require HP. Our results suggest that the relationship between WML and selective attention depends on the nature of the load and on the relationship between the relevant and irrelevant information.
This PDF is available to Subscribers Only