Oscillation Workshop (Oxford)

I had the privilege to get invited to speak at a meeting on brain oscillations at Oxford last week. It was a small (internal) meeting with some of the most exciting researchers in the field invited from all around the globe. Here are my thoughts and comments:

• Defining oscillations and naming them remains unresolved. What exactly is a gamma oscillation? Do I need narrow band peak in the spectrum to call it an oscillation? This gets even more confusing when comparing different species. such as for example when discussing theta oscillations in mice and men. I have written about this in my book Network Neuroscience, where I suggested that using Greek letters with arbitrary frequency cutoffs is more confusing than helpful. Similarly, at the meeting there was quite a consensus that oscillations should be defined by multiple features, including (perhaps) functional role, mechanisms, cellular and network substrate, etc.
• Dr. Norden Huang, inventor of the Hilbert-Huang Transform, gave an exciting  keynote lecture on how most time-series with periodic structure (especially in the brain) violates the assumptions of the Fourier transform (stationary, linear, etc). He has now developed a new method called the Holospectrum, definitely worth checking out if you are interested in avoiding artifacts such as harmonics in your analysis of the periodic structure of your series. He had some neat examples showing that, for example, if a sine wave is multiplied with white noise, spectral analysis cannot recover the sine-wave amplitude modulation despite it is clearly there in time-domain. 
• It was interesting to see that brain stimulation as a tool to demonstrate the causal role of oscillations in behavior has not yet fully reached cognitive neuroscience. The task is on us to convince the field that our techniques provide a conceptual advance and are as specific as we think they are.

Happy Monday,
Flavio