13 Aug 2009

Kristjan Korjus

Affiliated members Comments Off

Over the last five decades, complexity of the data in neuroscience has increased considerably: number of simultaneously recorded neurons has been doubled in approximately every 7 years and number of electrodes used by noninvasive technologies has increased from mere 2 to 256 (1).

The multidimensional nature and complexity of the data is a good indication that machine learning algorithms could be useful in this context, and indeed they have already been successfully used and will most likely only become more popular (2,3).

I will combine ideas of two work groups: BIIT  research group (machine learning) and prof. Talis Bachmann’s lab (cognitive neuroscience) (4,5) to one interdisciplinary approach in order to help answer neuroscientific problems. The working title of my PhD thesis is “Multidimensional electroencephalography (EEG) data analysis using machine learning algorithms”.

(1) Stevenson, I.H. and Kording, K.P., “How advances in neural recording affect data analysis,” Nature Neuroscience, vol. 14, no. 2, p. 139-142, 2011.

(2) Quiroga, R.Q. and Panzeri, S., “Extracting information from neuronal populations: information theory and decoding approaches,” Nature Reviews Neuroscience, vol. 10, p. 173-185, 2009.
(3) Haynes, J.D. and Rees, G., “Decoding mental states from brain activity in humans,” Nature Reviews Neuroscience, vol. 7, no. 7, p. 523-534, 2006.
Comments are closed.