This thesis used pupillometry to investigate whether pupils respond differently to faces that differ in familiarity. We aimed to see whether pupillometry measures cognitive processes involved in face processing, and whether it could be applied forensically. We started by evaluating three explanations for pupillary changes that occur when processing faces. The first was cognitive load (mental effort), because faces that have only been seen briefly are more difficult to recognise than well-known faces. The second was cognitive engagement (interest), because faces contain socially-important information. The third was memory strength (forensically applicable), as eyewitnesses have to recall a perpetrator's face in an attempt to identify them if they appear in a lineup. While pupillary responses reflected cognitive engagement to some extent, cognitive load best accounted for decreasing pupil sizes when learning new faces, and memory strength explained the pupillary changes seen in lineups. The theories all had some influence on pupil sizes, but their influence varied according to context, saliency, and the task at hand. Then we investigated whether pupillometry measured implicit recognition of a perpetrator in a lineup, and found that it did. Pupil sizes reflected memory strength in participants who believed their memory to be strong: there were differences in pupil sizes (between looking at the perpetrator and the distractors) in participants who identified him, but not in those who did not. The pupillary responses of participants who 'guessed' indicated that they were indeed guessing. There were no pupillary changes when the perpetrator was not in the lineup, even when participants misidentified a distractor. We concluded that pupillary responses are independent of explicit identification responses, and could be used forensically to support traditional measures of eyewitness identification and credibility.
|Publisher||University of Sussex|
|Source Sets||Ethos UK|
|Type||Electronic Thesis or Dissertation|
Page generated in 0.0061 seconds