Analysis of FMRI bold activation during the tower of London task using cortical pattern matching

We tested the hypothesis that novel cortical surface modeling approaches can permit more accurate localisation of functional deficits in schizophrenia, and more sharply defined activation foci. We created two average models of cerebral cortex from 3D structural magnetic resonance images (sMRI) of ten first-episode male patients with schizophrenia and ten age- and sex- matched controls. 3D functional magnetic resonance imaging (fMRI) BOLD activation data, obtained while subjects performed the 'Tower of London' task, was then mapped to both cortex models for comparison across diagnostic groups (patient and control) and between cortical alignment methods. The first average cortical model of cortex was created by identifying 17 salci on each hemisphere of each subject's brain. These sulci were geometrically averaged, and each subject's cortex warped into alignment with the average sulci. The second average cortical model of cortex was developed by aligning individual sMRIs to a template, intensity averaging, and extracting the cortex from the resulting average image, fMRI data showed increased activation with task difficulty in the control group in the left Brodmann's areas 44, 7, 37, and 18 and in the fight Brodmann's area 7. In the patient group, increased activation with task difficulty was seen in the left Brodmann's area 10 and in the right Brodmann's area 18. Significantly decreased activation with task difficulty was found in the left and fight Brodmann's area 9 of the patient group; this was not observed in the control group. Sulcal averaging and cortical data warping resulted in sharper localisation of the functional data when compared to the functional data mapped to the cortical model derived from an average intensity brain. Future comparisons will relate these functional differences to differences in cortical grey matter density in patients and controls.