Pupillary Dilation During Recognition Memory: Episodic Recovery or Goal Attainment?

A robust finding in eye tracking studies of memory is that successful recognition of studied items (i.e., old items) yields greater pupillary dilation compared to the successful recognition of unstudied items (i.e., new items), a finding termed the "pupil old/new effect". Given that the fundamental difference between recognizing old and new items is the presence of episodic content for old material, it is thought that the increased cognitive load incurred when episodic content is successfully retrieved drives the larger pupil dilation response (Goldinger & Papesh, 2012; Beatty, 1982; Kahneman, 1973). While the retrieval of episodic content may explain the pupil old/new effect, it is also possible that subjects are internally motivated to detect studied materials more so than unstudied materials. In other words, correctly recognizing old items is more valuable to subjects than correctly recognizing new items. Thus, the cognitive basis of the pupil old/new effect remains unclear because typical memory paradigms do not control for the motivational significance of recognition memory decisions. In order to de-confound episodic content and motivational significance, we used a performance-based incentive paradigm to manipulate subjects' motivation for recognizing old versus new materials. In separate test blocks, either “old” or "new" judgments were incentivized. If episodic content drives the pupil old/new effect, then dilation will remain larger for correctly recognized old items regardless of the incentive manipulation. If, however, the pupil old/new effect is not present when subjects are motivated to recognize new items, this suggests that pupillary dilation during recognition memory reflects goal attainment and not the recovery of episodic content.



The Effects of tDCS on Decision Bias and Pupillary Dilation During Recognition Memory

Recognition memory–the ability to distinguish novel from familiar stimuli–depends not just on our past experience (or lack thereof) with a stimulus, but also on our expectations given the environmental context in which it is encountered. Indeed, our expectations are known to influence our willingness to claim that something has been previously experience (i.e., response bias). Even so, virtually nothing is known about the cognitive and neural mechanisms that underlie adaptive, ‘on-the-fly’ changes in recognition bias. I will use an explicit cueing paradigm, which directly manipulates participants' expectations, in combination with non-invasive brain stimulation to test whether left lateral posterior parietal cortex (PPC) has a functional role in biasing recognition memory decisions and whether any changes in bias are reflected in pupillary dilation (an implicit measure of recognition). 


Sequential Dependencies in Metacognitive Confidence

Correlations between adjacent responses—known as sequential dependencies—in detection tasks have been well known to perception researchers for decades (e.g., Howarth, & Bulmer; 1956). However, sequential dependencies in metacognitive confidence ratings assessed during recognition memory have never–to the best of our knowledge–been reported in the literature. Thus, we are investigating whether sequential dependencies exist in recognition confidence judgments. We have found that (1) sequential dependencies exist in recognition confidence judgments, (2) these sequential dependencies persist when a perceptual task (gender identification) was interleaved between recognition trials, and (3) persist even when the intervening task was made to be very difficult. In fact, the subjects that had the difficult gender identification task showed greater sequential dependencies in their recognition confidence ratings than the subjects who experienced the easy task. Currently we are testing whether manipulating the difficulty of the perceptual task within subjects would moderate or amplify the effects of difficulty previously found between subjects. Secondly, we added confidence ratings to the gender identification task to test whether the effect of one confidence rating on another is domain specific or domain general.


Metacognition and Eye Movement Indices of Associative Recognition Memory

During the encoding phase of a memory task, subjects were asked to make predictions (termed judgments of learning, JOLs) about whether they would later remember each item during a subsequent testing phase. We found that items which received the highest JOLs (i.e., I will definitely remember this later) during encoding were later associated with quicker eye-movements.

Lisa A. Solinger