Maryam Vaziri-Pashkam

The paper investigates how artificial neural networks trained on motion-related tasks develop representations that align with brain activity and support object categorization using motion cues.

The paper shows that attentional bias in human visual cortex decreases with increasing target-distractor similarity, explained by tuning sharpening rather than gain increase.

The paper investigates whether visual cortex responses to graspable food items and tools are driven by shared behavioral affordances rather than distinct categories.

The paper investigates dual-task interference in the human brain using EEG, MVPA, and DDM, finding increased reaction times and reduced decoding accuracy in short SOA conditions.

The paper uses MVPA to show that dual-task interference in simulated driving disrupts information processing in the superior parietal lobe, correlating with slower reaction times.

The paper demonstrates that divisive normalization models better describe neural responses in the human visual cortex compared to weighted sum or average models, especially under attention conditions.

THINGS-data provides multimodal neuroimaging and behavioral datasets for studying object representations, enabling hypothesis testing and cross-modal analysis.

The paper investigates how motion and static visual cues are processed in the brain for object categorization, finding that both elicit robust category responses in visual cortex regions.

This paper compares transformation tolerance in human brain and CNNs, finding that the brain develops tolerance across feature changes while CNNs do not.

The paper identifies two brain networks with position-tolerant action representations, showing distinct relationships with objective movement measures and subjective action judgments.

The paper investigates how object identity and non-identity features are represented in the human ventral visual pathway using linear mapping functions across different transformations.

The paper compares fMRI responses from the human brain with 14 CNNs, finding partial correspondence in lower layers but not in higher layers or artificial objects.

The paper compares the coding strength of object identity and nonidentity features in human visual cortex and CNNs, finding increased identity representation and decreased nonidentity representation along the ventral visual pathway.

This study investigates dual-task interference in a simulated driving environment, finding that tasks are processed in a partial-parallel manner.