Poster A81, Saturday, March 24, 1:30–3:30 pm, Exhibit Hall C
Novelty enhances the reliability and timing consistency of neuronal source response
Guang Ouyang1, Yunqing Hua1, Changsong Zhou2, Akaysha Tang1; 1Laboratory of Neuroscience for Education, Faculty of Education, The University of Hong Kong, 2Department of Physics, Institute of Computational and Theoretical Studies, Hong Kong Baptist University
Brain response to unexpected or novel environmental changes is critical for learning and survival, therefore should be reliable and consistent across instants. While the P300, a large-amplitude neural response to low-probability stimuli, has been extensively investigated using scalp-recorded event-related potentials (ERPs), its neural origin, reliability and variability at single trial level are under-explored. Applying second-order blind identification (SOBI) algorithm to continuous data collected during a color visual oddball task, we characterized novelty responses in two contrasting neuronal sources: an early visual source (peak latency 111±16 ms) localized to occipital gyrus and a late P300 source localized to a network of frontal, occipital and temporal lobe structures. Single-trial analysis of single subject source data showed that 1) both sources showed significantly greater ERP amplitudes to the rare than to the frequent stimuli (50:200); (2) In only the P300 source, single trial ERP response to low-probability stimuli is significantly more reliable (fewer misses) and consistent (lower standard deviation of latency) than to the frequent stimuli; 3) In 9 out of 11 participants, the single-trial P300 peak latencies predicted reaction times of button-press to low-probability stimuli; 4) The early visual and late P300 sources are correlated in peak latency during trials of low- but not high-probability stimuli. Methodologically, these results demonstrate a new capacity for estimating single-trial response characteristics in functionally specific brain regions. Scientifically, these results suggest that novelty detection is a significant neural computation performed at both early sensory processing stage and subsequent evaluation and action generation stage.
Topic Area: METHODS: Neuroimaging