Description

The neural mechanisms involved in recognizing and classifying the thousands of complex visual objects we encounter every day are not fully understood. Previous researchers have examined the processing of other categories of stimuli (e.g., visual words, auditory words, and sign language stimuli) in so-called “megastudies” by including a large number of participants and a huge sample of stimuli. The current study brings the megastudy approach to the investigation of the temporal dynamics of visual object recognition using event-related potentials (ERPs). The primary goal was to engage meaning-based neural representations as participants viewed color pictures of 1000 different objects. EEG was collected from 52 participants engaged in a go/no-go semantic categorization task where they responded with a button press (“go”) to occasional objects (10%) that were “kitchen” items but withheld responding (“no go”) to the remaining items (90%). Here we report on differences between ERPs to common real objects and non-real objects (e.g., abstract shapes, fictitious things, and manipulated images of real objects), objects rated high and low in familiarity, and objects that depicted animals and tools. We found that animals elicited a greater anterior negativity while tools elicited a greater posterior negativity during the N300 and N400 time epoch. Also, ERPs to non-objects generated a larger anteriorly distributed negativity in the N400 epoch than familiar objects. Finally, objects rated as less familiar produced more negative N300 and N400 activity than more familiar objects. We also contrasted these comparisons with those reported last year using the same stimuli but a different task (object decision). In that study there were similar ERP effects to those reported above for object familiarity and real vs. non-real objects. However, in the animal versus tool comparison only the posterior N300/N400 effect was seen in object decision suggesting that the anterior negativity in the current study reflects the task-dependent semantic demands of categorization that are not accessed during object decision.