A personalized history of EEG–fMRI integration

Abstract

In this personalized review, the history of EEG recorded simultaneously with functional MRI (EEG–fMRI) is summarized. A brief overview is given of the hardware development followed by a discussion of EEG–fMRI applications. The technique's development was clinically motivated in the context of epilepsy. Accordingly, the evolution of studies trying to identify with EEG–fMRI the origin of individual epileptiform discharges up to those revealing general mechanisms of epilepsy syndromes is sketched. In epilepsy centers experienced with the methodology, EEG–fMRI has found its place in the presurgical evaluation of patients. In cognitive neuroscience, the multimodal technique has significantly contributed to the understanding of phenomena of the resting state and neuronal oscillations. While most of the studies discussed relate EEG to fMRI by means of prediction, the development of forward models facilitating the symmetrical ‘fusion’ of EEG and fMRI data is the subject of current neuro-computational research. Recently, intracranial EEG has been safely recorded during (functional) MRI broadly extending the perspectives for epilepsy and research into neurovascular coupling. EEG–fMRI has evolved into a mature, generally accessible and in principle easily applicable technique, which is a great achievement. Because this at the same time bears the risk of unreflected use, EEG–fMRI safety issues are also highlighted.

Introduction

Peter Bandettini asked all authors of this Special Issue to make contributions from the “unique perspective of each author”. This is why the following account in parts will be biased toward the European history of EEG–fMRI integration, specifically a “London perspective” (Hamandi et al., 2004), the one which I was most closely involved with. A more balanced view can be obtained by studying review articles on the topic which will show that many more groups contributed greatly to the field, such as those around John Ives, Franz Schmitt, Steven Warach and Donald Schomer in Boston/U.S.A. (Ives et al., 1993, Warach et al., 1996), John Archer, David Abbott and Graeme Jackson in Melbourne/Australia (Archer et al., 2003a, Archer et al., 2003b), Margitta Seeck, Christoph Michel and Theodor Landis in Geneva/Switzerland (Seeck et al., 1998), Jean Gotman and colleagues in Montréal/Canada (Al-Asmi et al., 2003, Benar et al., 2002, Benar et al., 2003), Alexander Hoffmann, Lorenz Jäger and Maximilian Reiser in Munich/Germany (Hoffmann et al., 2000, Jäger et al., 2002) — just to name a few.

On March 2nd, 2002, Louis Lemieux and Robert Turner held the ‘First Workshop on EEG–fMRI’ at Queen Square in London with David Fish (Institute of Neurology, UCL, UK), Georgio Bonmassar (Hardvard, U.S.A.), John Stern (UCLA, USA), Afraim Salek-Haddadi (Institute of Neurology, UK), Walter Freeman (Berkeley, USA), Arno Villringer (Charité, Germany), Jean Gotman (Montreal Neurological Institute, Canada) and Fabio Babiloni (Roma 1, Italy) as the speakers. I had the opportunity to attend this in retrospect historical meeting accompanying Karsten Krakow. He had completed his PhD at UCL as the first medical fellow (under David Fish) acquiring EEG–fMRI at the National Society for Epilepsy (now called The Epilepsy Society, Chalfont St. Peter, UK) from Queen Square (The National Hospital for Neurology and Neurosurgery, UCL, UK) epilepsy patients (Krakow et al., 1999) with Philip Allen's MR-compatible EEG system (Hamandi et al., 2004, Krakow et al., 1999). Karsten Krakow after his PhD had moved to the Department of Neurology at the Goethe University in Frankfurt (Germany), where I met him starting my fellowship in neurology. It was at that symposium that apart from the speakers I had the opportunity to meet in person some of the “London EEG–fMRI pioneers” including Phil Allen, Oliver Josephs and Mark Symms.

At the inception of EEG–fMRI, advances on the technical as well as the analysis side were tremendous and went hand in hand with one another (first part of this review), while later on, when the first major technical hurdles had been taken and good commercial hard- and software were available, scientific applications and analysis strategies could advance independently of the engineering side of matters (second part). In my opinion, the future progress of EEG–fMRI integration again will depend on further technical advances required to extend current boundaries of EEG–fMRI partly posed by physics and safety limitations (part three).