Better method breeds better science
By Dr Aina Puce
Image credit: Science by Joel Filipe. CC0 Public Domain via Unsplash.
It has become obvious over the last couple of decades that available opportunities to train in EEG/MEG-based methods have lagged behind the existing demand to learn. As a journal editor, grant review panel member, and EEG/MEG course teacher, the absence of material that integrated the underlying basis and principles of both EEG and MEG (read this freely available chapter) side-by-side was surprising – these are, after all, two complimentary methods that measure neural activity directly.
Often, investigators understand fMRI, but not EEG and MEG. Alternatively, an investigator might be an expert in EEG and not MEG, or vice versa, and not understand fMRI. In reading manuscripts and grant reviews, it may be methods do not get adequate scrutiny. Reviewers often have no direct experience or no background knowledge of the methodology, which can produce the “but the ideas are good” argument, ignoring the notion that the proposed methods will not answer the scientific question being posed. Appropriate and sound method should be the backbone of any study, given the current debate on reproducibility in science. Without this, science becomes a house of cards, as layers of potentially erroneous or uninterpretable findings are built upon one another.
Shifting sands in cognitive and social neuroscience. As scientific practice becomes less reductionist, training scientists with a basic understanding of multiple methods has become increasingly important. Continued progress in scientific disciplines using in vivo studies of brain activity and behavior relies on this premise. Early studies in the 1960-80s used electroencephalography (EEG), and subsequently magnetoencephalography (MEG), allowing the time course of brain activity to be visualized. From the latter part of the 20th century, neuroimaging methods such as positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) became very popular, although EEG/MEG studies continued during this time. The popularity of fMRI has ensured a very well-stocked cadre of investigators who are experts in this methodology. At the same time, the relative popularity of EEG/MEG has dropped somewhat, as cognitive and social neuroscientists sought training in MRI-based methods. It is clear that EEG/MEG are important for studies of cognitive and social neuroscience, where timing of neural activity and putative sources of activity in brain networks can be studied with millisecond accuracy.
Benefits to the community. A better understanding of methods produces better science in multiple ways. First, understanding multiple methods allows experimental questions to be answered optimally – using methods appropriate for each question. Second, wider surveys of the existing literature can better interpret new data. Third, better scientific theories can be constructed if they are based on data (and literature) that have not been methodologically constrained, e.g. incorporating findings from fMRI, EEG and MEG studies. Finally, scientific dexterity enables critical, but fair, peer review, as discerning reviewers understand the proposed methods and their limitations. In short, science benefits overall.
Dr Aina Puce is a social neuroscientist with research interests in the brain bases of human non-verbal communication. Her studies in basic and clinical human neuroscience have used scalp and intracranial EEG, and functional MRI methods. Her formal training was in biophysics and functional brain mapping/neurophysiology.
Dr Puce is the co-author of MEG-EEG Primer, along with Dr Riitta Hari, which is available in print and online from Oxford University Press. To date no one has attempted to integrate the two methods of MEG and EEG in a single tome, but it is something the authors felt was crucial for a more complete understanding of non-invasive human neurophysiology.
For other articles, visit our Article Archive.