Proceedings of EEG Spring School at King’s College London

by Glenn Kitsune, PhD1 & Grainne McLoughlin, PhD2
1Post-Doctoral Study Manager and Course Leader for EEG Spring School, Department of Psychological Medicine, Institute of Psychiatry, Psychology and Neuroscience, King’s College London
2Lecturer in Cognitive Neuroscience and Director of EEG Studies, Social Genetic Developmental Psychiatry Centre, Institute of Psychiatry, Psychology and Neuroscience, King’s College London

Organizers.jpg

Workshop Organisers: Freya Rumball, PhD (Deputy Course Leader), Grainne McLoughlin, PhD (Director) and Glenn Kitsune, PhD (Course Leader)

The Spring School in EEG analysis is designed to provide primary training in both the collection and analysis of EEG data. The course provides a balance of both theoretical knowledge and practical data processing skills using BrainVision Analyzer 2.1 software. The course was set up to contribute basic to intermediate training in EEG to keep pace with the ever increasing demands for use of EEG methods.

The initial workshop was organised into four days. Day One consisted of practical training in EEG data collection in small groups. There were a limited number of places available for this unique one-day training session. Attendees got to spend time with our expert EEG researchers in our dedicated SGDP EEG lab equipped with Brain Products’ actiCAP system. During these sessions delegates conducted a realistic EEG session with live volunteers where they could learn tips and techniques and receive hands-on training on how to get the best possible EEG data from their research participants.

Day Two focused on basic understanding of EEG data and data pre-processing, including the use of Independent Component Analysis (ICA) for ocular artefact correction. Day Three focused on choosing optimal methodology and research design, and conducting quantitative event-related potential (ERP) and spectral analysis.

Day Four was a ‘conference’ day, with leading EEG researchers from across the UK, Europe and US invited to give talks on their research to highlight the advanced analysis possibilities available with EEG data.

The day started with a talk by Dr. Charlotte Tye, postdoctoral fellow at King’s College London (KCL), on the transition from spontaneous pre-stimulus EEG activity to more ordered post-stimulus oscillations. Dr. Tye’s innovative work shows that activity occurring during the pre-stimulus period plays a significant role in modulating post-stimulus behavioural and neurophysiological responses. This has implications for our understanding of individual reactions and adaptations to the environment both in typical and atypical development.

This was followed by a talk by Dr. Judith Nottage, also a postdoc at KCL, who discussed fast-occurring neural oscillations, specifically the gamma band (> 30 Hz). Human and animal studies have shown that these oscillations increase the gain of neural signals in response to attention, and are associated with active brain states and correlated with the fMRI BOLD response. Gamma is thought to be altered in human pathological states such as schizophrenia and autism. The contamination of scalp EEG recordings in the gamma range with artefacts has made extracting the genuine neural signal challenging. In addition to 50 Hz noise from alternating current, the muscles of the eyes and scalp generate electrical signals in the gamma band. Dr. Nottage gave a comprehensive review of approaches to dealing with these artefacts, including covering her own proven and useful work to uncover the true gamma signal.

Drs. Gráinne McLoughlin (Lecturer in Cognitive Neuroscience at KCL) and Christopher Saville (postdoc at Universities of Bangor and Freiburg) discussed their different, yet complementary, approaches to maximise the signal-to-noise ratio of evoked brain activity. Brain activity is usually averaged across many trials to compute event-related potentials (ERPs) and oscillations (EROs). This rests on the assumption that the underlying signal does not vary across trials and all that does vary is irrelevant noise. While this simplifying assumption has led to great progress for the field, it is undoubtedly incorrect. Both of these talks presented methods for identifying oscillatory signals and ERPs in individual trials.

While Dr. Saville primarily focused on P3b activity, Dr. McLoughlin focused on oscillatory activity in the alpha and theta ranges. Dr. McLoughlin discussed the application of independent component analysis (ICA) to EEG data, focusing in particular on the extraction of trial-by-trial phase and amplitude measures of the associated EROs. Neural oscillations are a fundamental mechanism for enabling coordinated activity during normal brain functioning. Dr. McLoughlin highlighted in her talk that deficits in neural oscillations may be sensitive and specific to the pathophysiology of a number of psychiatric disorders and may represent functional disconnection between and within cortical areas of the brain.

The talks were rounded off by Dr. Jason Palmer from the Swartz Center for Computational Neuroscience at University of California, San Diego. In line with the advanced methods discussions, Dr. Palmer described the standard ICA model of EEG generation, and considered the extent to which this model is valid in practice. Dr. Palmer’s analysis shows that the model is in fact very good in many circumstances, with plausible underlying physiological mechanisms. He considered two circumstances in which the standard model breaks down: the case of more components than channels, and the case of dependent (neural) components, and proposed an extension of the standard ICA model (AMICA) to accommodate these cases. Finally he described statistical methods including mutual information for quantifying the significance of independence/dependence of sources.

The course was oversubscribed and delegates who managed to gain a place travelled from all over the UK and Europe to attend the course. We received extremely positive and useful feedback with over 4 out of 5 on rating measures, including ‘enjoyability’ and ‘expertise of tutors’. Quotes from the students included, “All talks and practical sessions were really good. Particularly having a large number of knowledgeable demonstrators available to provide advice during practical sessions was very good” and, “Teaching very clear, very useful in understanding the basics and getting to a point where I can do this independently”.

KCL EEG Spring School 2016As we want to continue to adapt and develop this course to best meet the needs of the attendees, in response to feedback subsequent courses will now include an additional day for further practical analysis training. Given the apparent demand, the next course will run in spring 2016. For further information, please contact Please contact Grainne McLoughlin (grainne.mcloughlin@kcl.ac.uk).

Brain Products kindly sponsored the event and provided software and licenses for the practical sessions, without which this essential training in basic EEG data collection and analysis would not have been possible. We are very grateful to Brain Vision UK, in particular Mary Tighe, for helping us with organising and conducting the SGDP EEG School – we look forward to our future collaboration on this event. We also want to thank all of our staff who worked on the EEG course with us, in particular, Dr Freya Rumball.

©Brain Products GmbH 2015

Print Friendly, PDF & Email
4 0