Grey matter volumetric changes related to recovery from hand paresis after cortical sensorimotor stroke.

Preclinical studies using animal models have shown that grey matter plasticity in both perilesional and distant neural networks contributes to behavioural recovery of sensorimotor functions after ischaemic cortical stroke. Whether such morphological changes can be detected after human cortical stroke is not yet known, but this would be essential to better understand post-stroke brain architecture and its impact on recovery. Using serial behavioural and high-resolution magnetic resonance imaging (MRI) measurements, we tracked recovery of dexterous hand function in 28 patients with ischaemic stroke involving the primary sensorimotor cortices. We were able to classify three recovery subgroups (fast, slow, and poor) using response feature analysis of individual recovery curves. To detect areas with significant longitudinal grey matter volume (GMV) change, we performed tensor-based morphometry of MRI data acquired in the subacute phase, i.e. after the stage compromised by acute oedema and inflammation. We found significant GMV expansion in the perilesional premotor cortex, ipsilesional mediodorsal thalamus, and caudate nucleus, and GMV contraction in the contralesional cerebellum. According to an interaction model, patients with fast recovery had more perilesional than subcortical expansion, whereas the contrary was true for patients with impaired recovery. Also, there were significant voxel-wise correlations between motor performance and ipsilesional GMV contraction in the posterior parietal lobes and expansion in dorsolateral prefrontal cortex. In sum, perilesional GMV expansion is associated with successful recovery after cortical stroke, possibly reflecting the restructuring of local cortical networks. Distant changes within the prefrontal-striato-thalamic network are related to impaired recovery, probably indicating higher demands on cognitive control of motor behaviour.

Artificial neural network Radon inversion for image reconstruction.

Image reconstruction techniques are essential to computer tomography. Algorithms such as filtered backprojection (FBP) or algebraic techniques are most frequently used. This paper presents an attempt to apply a feed-forward back-propagation supervised artificial neural network (BPN) to tomographic image reconstruction, specifically to positron emission tomography (PET). The main result is that the network trained with Gaussian test images proved to be successful at reconstructing images from projection sets derived from arbitrary objects. Additional results relate to the design of the network and the full width at half maximum (FWHM) of the Gaussians in the training sets. First, the optimal number of nodes in the middle layer is about an order of magnitude less than the number of input or output nodes. Second, the number of iterations required to achieve a required training set tolerance appeared to decrease exponentially with the number of nodes in the middle layer. Finally, for training sets containing Gaussians of a single width, the optimal accuracy of reconstructing the control set is obtained with a FWHM of three pixels. Intended to explore feasibility, the BPN presented in the following does not provide reconstruction accuracy adequate for immediate application to PET. However, the trained network does reconstruct general images independent of the data with which it was trained. Proposed in the concluding section are several possible refinements that should permit the development of a network capable of fast reconstruction of three-dimensional images from the discrete, noisy projection data characteristic of PET.