MR mapping of neuronal activation. A study of acquisition sequences

Abstract: Functional magnetic resonance investigation of neuronal activation is increasingly applied in clinical practice and in research in the neurosciences. Given the quantitative non-reproducibility of published results (numbers of pixels activated relative signal intensity) we systematically analysed the acquisition parameters to determine their effect on activation mapping, in the present study activation maps of the motor and visual cortex. We investigated 21 healthy volunteers studying the following parameters: echo time, slice thickness, flip angle, field homogeneity, spatial orientation of the slices. Data were analysed in parts. Firstly, we calculated the percentage variation of the map (PCSI, PerCentage Signal Image) of images. With TASK defined as the matrix obtained from the average images acquired during activity (excluding the first two in each cycle) and REST the matrix obtained from the average images acquired during rest cycles, the PCSI map was defined as follows: PCSI = (TASK-REST) / REST The PCSI maps were characterized by a high level of noise: alongside variations in intensity correlated to stimulation, there were also many chance contributions. Secondly, data were analysed statistically in terms of the correlation between temporal trend of the signal of each pixel in the image and the temporal trend of the paradigm examined (Signal intensity: rest: = 0, activity = 1). The pixels whose analysis was significant were included in the activation map only if they were in clusters of at least three and excluded if isolated. An average signal increase was observed with the echo time: variations in echo time from 20 to 40 ms were accompanied by a PCSI increase of 80%. Reducing the slice thickness from 12 to 4 mm we increased PCSI by a factor of two. In our experience an acquisition matrix 128x128, reconstructed 256x256, with a slice thickness of 3 or 4 mm is a good compromise between spatial resolution, signal-to-noise ratio and acquisition time for these studies The pixels corresponding to the drainage macrovasculature usually showed a higher average PCSI (>10%) than that observed with pixels activated by the grey matter In addition, the average and peak PCSI increase was much more evident for the macrovasculature than for tissues. The number of pixels malting up the activation map grows with the flip angle, probably because the signal-to-noise ratio improves as the flip angle increases. Studying the effects of inhomogeneity we found that the width at half height of 20Hz for the water peak was sufficient to calculate the maps and can be obtained automatically in a few minutes. For motor task neuronal activation, we found that optimisation of the magnetic field increased the number of pixels recognised as activating, whereas the average value of intensity of the pixels activated was unchanged for visual stimulation tests, we found that the number of pixels activated remained the same after shimming, whereas the average intensity of pixels activated doubled. We therefore devised an optimal protocol for the clinical application of the fMRI technique using a 1.5 T system: TE = 40 ms, TR = 60 ms, flip angle = 25 degrees, slice thickness = 4 mm, acquisition matrix = 128x128, reconstruction matrix = 256x256, compensation for flow artefacts, contrast enhancement technique = T1.