will take place on Friday, June 28, 2019 from 15:00 to 16:00 hours in CBBM, Ground Floor, Seminar Room B1/B2.
Host: Prof. Dr. Jonas Obleser
Institute of Psychology I
University of Lübeck
Abstract
Moment-to-moment variability of neural activity is increasingly recognized as essential for cognitive ability. One fruitful approach to quantifying neural variability in EEG is by estimating its complexity on shorter and longer time-scales through multi-scale entropy (MSE). Traditional MSE, however, suffers from both the need for long, continuous data for estimation and a lack of temporal specificity, limiting its use for studying fast cognitive processes such as decision making.
In this talk, I will present a recently developed, modified MSE (mMSE) estimation method that overcomes both these limitations and shows promise to reveal links between neural variability and cognitive function. Specifically, instead of employing continuous EEG data, mMSE estimates entropy on concatenated data segments taken across trials using a sliding time window, allowing the tracking of neural variability over time as the cognitive process unfolds. In addition, estimating entropy on concatenated data segments allows probing entropy on longer time-scales that would normally remain unresolvable within the limited data available in a single trial.
I will showcase mMSE using EEG data from a previously published study investigating the link between spectral power and strategic shifts in decision bias that participants applied to maximize reward. Strikingly, we find that mMSE on longer time-scales reliably predicts decision bias shifts across participants. This effect cannot be explained by either spectral power in any frequency band or traditional MSE, suggesting that mMSE uniquely captures variability that is specifically relevant for behaviour.
I will end the talk with an outlook of mMSE’s use for the study of neural variability in different aspects of human cognition.
