A human single-neuron dataset for object recognition

Area of Science:

• Neuroscience
• Cognitive Science
• Computer Vision

Background:

• Object recognition is crucial for human interaction and interpretation of the environment.
• The amygdala and hippocampus are key brain regions involved in encoding and retrieving visual information for object recognition.

Purpose of the Study:

• To record single-neuron activity in the human amygdala and hippocampus during an object recognition task.
• To analyze the neural correlates of object recognition using naturalistic stimuli.
• To provide a dataset for training and validating computational models of object recognition.

Main Methods:

• Single-neuron activity was recorded from neurosurgical epilepsy patients.
• A one-back task was performed using naturalistic object stimuli from ImageNet and Microsoft COCO datasets.
• Data were collected from 1204 neurons using ImageNet stimuli and 512 neurons using Microsoft COCO stimuli.

Main Results:

• The study generated an extensive dataset of human neural activity during object recognition.
• The dataset offers high spatial and temporal resolution for analyzing neural correlates.
• The findings will aid in understanding the neural basis of object recognition and computational modeling.

Conclusions:

• The recorded neural data provides a valuable resource for studying object recognition in the human brain.
• This research facilitates the development and validation of computational models for visual processing.
• The findings contribute to a deeper understanding of how the human brain recognizes objects.