Increasing EEG Generalization

The following research was completed at the Attention Perception and Performance Laboratory in the Department of Psychology, within the Faculty of Science, at the University of Alberta

The following body of work approaches the complex problem of generalization in typical neuroscience research from different angles, an example of the multi-modal conceptualization I bring to problem solving.

No jittering with Latte Panda, transitioning EEG experimentation out of the lab

Electroencephalography (EEG) is typically conducted in a highly controlled laboratory setting. However, this limits the generalizability of results to real-world situations.

To further increase the portability of EEG experiments, we explored the use of a Latte Panda, a compact and relatively inexpensive Windows PC board, in the collection of EEG data. To assess data quality between a Latte Panda and Windows laptop, we compared the P3 and MMN waveforms elicited during an auditory oddball task. In this task, participants listened to either high-pitched or low-pitched tones and pressed a button each time a high tone is presented, with low tones being played 80% of the time and high tones only 20%. The MMN waveform is a negative deflection following the presentation of the rare tone whereas the P3 waveform is a positive deflection following the rare tone. Our results suggest that the Latte Panda can serve as a reliable replacement to a laptop regarding EEG data collection. Having established reliability of the Latte Panda, we then had participants engaged in an auditory oddball task while stationary cycling. Participants were fitted with a self-contained EEG backpack, and completed an auditory oddball task, while either sitting and biking, inside a Faraday chamber.

This serves to validate the next step towards taking experimentation into more generalizable contexts. Currently, results suggest that data is comparable between the biking and sitting conditions. Such results will allow for more portable, and affordable, EEG experimentation.

Real brains in virtual environments: An investigation of depth processing using a novel depth Oddball task

in collaboration with IBM Virtual Reality and Game Design within the Education & Cognitive Sciences Department

Electroencephalography (EEG) research is typically conducted in a highly controlled laboratory setting. However, this typically limits the generalizability of results to real-world situations. Contemporary research has shown that more portable means of stimulus presentation can yield results comparable to EEG traditional methods.

Alternatively, we are exploring the use of virtual reality for presenting stimulus within an EEG study. This may serve to characterized brain states in novel or otherwise inaccessible research environments. In the present study we used an HTC Vive head mounted display (HMD) within a Faraday chamber to assess brain states during a novel depth based P3. For this task, stationary participants responded to either near or far (size-matched) target orbs within a virtual environment. Standard orbs were presented 80% of the time and target orbs only 20%. Typical oddball task waveforms were analysed for each condition. Including the MMN and P3 waveforms, a negative and a positive deflection respectively, elicited following target orb presentation. Additionally, horizontal electrooculography (EOG) was fitted to measure eye convergence and divergence and was shown to be related to condition and orb type.

Further research will investigate the effects of objects at various depth planes in relation to brain states. Additionally, informative cues may be used to elicit the construction of object files. Currently, results suggest that virtual reality can serve as an valid alternate means of stimulus presentation in novel or otherwise inaccessible environments.

Thank you to all those that contributed to and supported this research, you can find the lab website here.

Google Sites

Report abuse