Zihan Bai
@zihanbai.bsky.social
PhD student at NYU working with Sebastian Michelmann and Daryl Fougnie. I study memory. I bring the past to the present.
(iii) Strikingly, two trials before the target appeared, predictive viewing of the first of two pairmates foreshadowed upcoming discrimination success, indicating predictive recall of diagnostic features.
September 2, 2026 at 11:34 PM
(iii) Strikingly, two trials before the target appeared, predictive viewing of the first of two pairmates foreshadowed upcoming discrimination success, indicating predictive recall of diagnostic features.
(ii) At retrieval, recapitulation of the target’s gaze pattern from encoding (pattern completion) marked an item’s discrimination success.
September 2, 2026 at 11:34 PM
(ii) At retrieval, recapitulation of the target’s gaze pattern from encoding (pattern completion) marked an item’s discrimination success.
We then identified three hallmark computations of EM that supported WM discrimination: (i) At encoding, distinctive viewing of pairmates predicted subsequent discrimination success (pattern separation).
September 2, 2026 at 11:34 PM
We then identified three hallmark computations of EM that supported WM discrimination: (i) At encoding, distinctive viewing of pairmates predicted subsequent discrimination success (pattern separation).
Tracking trial-by-trial eye movements, we first confirmed that gaze patterns during image viewing carry image-specific information at sufficient resolution: gaze pattern similarity tracked image similarity within and across tasks.
September 2, 2026 at 11:34 PM
Tracking trial-by-trial eye movements, we first confirmed that gaze patterns during image viewing carry image-specific information at sufficient resolution: gaze pattern similarity tracked image similarity within and across tasks.
We found a striking behavioral benefit of EM that was confined to discriminating similar items. This benefit was also modulated by encoding condition: items that had been compared to each other had the strongest benefit, while items shown in isolation showed a weaker advantage over novel pairmates.
September 2, 2026 at 11:34 PM
We found a striking behavioral benefit of EM that was confined to discriminating similar items. This benefit was also modulated by encoding condition: items that had been compared to each other had the strongest benefit, while items shown in isolation showed a weaker advantage over novel pairmates.
A one-back phase with low WM load served as an incidental encoding phase.
In a subsequent two-back phase with high WM load, some of the pairs from encoding reappeared, allowing us to test whether episodic representations of individual object images support their comparison in working memory.
In a subsequent two-back phase with high WM load, some of the pairs from encoding reappeared, allowing us to test whether episodic representations of individual object images support their comparison in working memory.
September 2, 2026 at 11:34 PM
A one-back phase with low WM load served as an incidental encoding phase.
In a subsequent two-back phase with high WM load, some of the pairs from encoding reappeared, allowing us to test whether episodic representations of individual object images support their comparison in working memory.
In a subsequent two-back phase with high WM load, some of the pairs from encoding reappeared, allowing us to test whether episodic representations of individual object images support their comparison in working memory.