Mikko Uusitalo

Brain Research Unit
Low Temperature Laboratory

Dissertation for the degree of Doctor of Technology presented at Helsinki University of Technology on June 9th, 1997
Opponent was Professor Michael Scherg, University of Heidelberg.

Pictures from the occasion


The thesis is based on 6 publications The first part of this thesis [papers 1 and 2] presents a mathematical analysis and a practical application of the signal-space projection (SSP) method for separating MEG into components. Both distributed and well-localized neuronal sources can be characterized. This method is now in use in several laboratories. With the component separation of SSP it is possible, e.g., to remove artefacts, to enhance the signal-to-noise ratio, and to identify weak sources hidden by stronger ones. The angle theta between component vectors in signal space measures the separation between the components. Errors in source parameters identified on the basis of data from which known source contributions have been removed, were shown to be inversely proportional to the signal-to-noise ratio and sin(theta), where theta is the angle between the modeled and the known source components. Application of signal-space projections in the frequency domain provides information on the frequency content of the signal components. SSP will likely remain as a useful standard tool.

In paper 3 a dynamic organization of the human visual system was revealed by measuring activation trace lifetimes in areas activated by a checkerboard stimulus. Each responding cortical area supported a memory function with a well-defined lifetime. The areas fell into two subgroups, the first in the occipital lobe with lifetimes ranging from 0.1 to 0.6 s, and the second in temporal, parietal, and frontal areas with lifetimes ranging from 7 to 30 s. Also, within each group the areas responding later tended to have longer lifetimes. In [paper 4], the average lifetime of 0.8 s in the motion-specific area V5 is in agreement with the assumed level of V5 in the hierarchy of visual information processing. These experiments open new possibilities for studying visual sensory memory.

In contrast to a recent fMRI study, equally strong V5 activation in both dyslexics and control subjects was detected in paper 5, with a trend for longer latency in dyslexics. The lack of fMRI signals in the V5 cortex of dyslexics may in part depend on methodology. This emphasizes the complementary properties of different functional imaging techniques. New comparative fMRI studies would be interesting. Further experiments are needed to clarify and characterize the deficits of dyslexics in sensory processing.

Responses to stationary and rotating radial gratings were compared in paper 6. Rotations lasting 1 s evoked sustained activity in the V5 complex region, on average 1.6 cm anterior to activation produced by linear motion stimuli in [P4]. This agrees with the anatomical organization of monkey visual cortex, in which area MST is located anterior to area MT. The human V5 complex showed a transient signal component 100 - 130 ms after onsets of both rotating and stationary stimuli, suggesting that the V5 complex responds to transient changes in the visual environment in addition to stimulus movement. Additional motion-related sustained activation was identified in the vicinity of the central sulcus.