Search

Purpose: Clinical outcomes vary considerably among individuals with vessel occlusion of the posterior circulation. In the present study we evaluated machine learning algorithms in their ability to discriminate between favourable and unfavourable outcomes in patients with endovascular treatment of acute ischaemic stroke of the posterior circulation., Methods: This retrospective study evaluated three algorithms (generalised linear model, K-nearest neighbour and random forest) to predict functional outcomes at dismissal of 30 patients with acute occlusion of the basilar artery who were treated with thrombectomy. Input variables encompassed baseline as well as peri and postprocedural data. Favourable outcome was defined as a modified Rankin scale score of 0-2 and unfavourable outcome was defined as a modified Rankin scale score of 3-6. The performance of the algorithms was assessed with the area under the receiver operating curve and with confusion matrixes., Results: Successful reperfusion was achieved in 83%, with 30% of the patients having a favourable outcome. The area under the curve was 0.93 for the random forest model, 0.86 for the K-nearest neighbour model and 0.78 for the generalised linear model. The accuracy was 0.69 for the generalised linear model and 0.84 for the random forest and the K nearest neighbour models., Conclusion: Favourable and unfavourable outcomes at dismissal of patients with acute ischaemic stroke of the posterior circulation can be predicted immediately after the follow-up non-enhanced computed tomography using machine learning.

Purpose: Functional magnetic resonance imaging (fMRI) localizes Broca's area (B) and Wernicke's area (W) and the hemisphere dominant for language. In clinical fMRI, adapting the stimulation paradigms to each patient's individual cognitive capacity is crucial for diagnostic success. To interpret clinical fMRI findings correctly, we studied the effect of varying frequency and number of stimuli on functional localization, determination of language dominance and BOLD signals., Materials and Methods: Ten volunteers (VP) were investigated at 1.5 Tesla during visually triggered sentence generation using a standardized block design. In four different measurements, the stimuli were presented to each VP with frequencies of 1/1 s, (1/2) s, (1/3) s and (1/6) s., Results: The functional localizations and the correlations of the measured BOLD signals to the applied hemodynamic reference function (r) were almost independent from frequency and number of the stimuli in both hemispheres, whereas the relative BOLD signal changes (DeltaS) in B and W increased with the stimulation rate, which also changed the lateralization indices. The strongest BOLD activations were achieved with the highest stimulation rate or with the maximum language production task, respectively., Conclusion: The adaptation of language paradigms necessary in clinical fMRI does not alter the functional localizations but changes the BOLD signals and language lateralization which should not be attributed to the underlying brain pathology.

Preoperative functional magnetic resonance imaging (fMRI) localizes the primary motor and somatosensory cortex in relation to rolandic brain tumors and determines plastic cortical reorganization. Functional landmarks help to assess the indication for surgery and to plan for safer surgical procedures that protect the functional cortex during resection even when morphologic landmarks are no longer identifiable on anatomic images. Despite its successful application, preoperative fMRI has not yet reached the status of an established clinical diagnostic procedure since special stimulation systems, standardized fMRI protocols and medically approved software are still lacking. Following a brief review of the image display of the functional and morphologic anatomy, the different indications for preoperative fMRI in patients with rolandic brain tumors are presented. A robust preoperative protocol enables clinical MR units with magnetic field strengths of 1.0 Tesla or higher to perform reliable fMRI during contralateral hand movements. Optimized investigation strategies and stimulation modalities are proposed for patients with rolandic tumors distant from the cortical hand representation, for patients with preexisting sensorimotor deficits and for patients with poor compliance. Representative cases illustrate the clinical application. Possibilities and limitations of preoperative fMRI are presented and discussed.

The human primary sensorimotor cortex was investigated for somatotopic organization during motor imagery (IM) which was compared to motor execution (EM). Block designed BOLD (blood oxygen level dependent)-functional magnetic resonance imaging at 1.5 Tesla was applied in 14 right handed volunteers during imagined and executed tongue, finger and toe movements. BOLD-clusters were assessed for anatomically correct sensorimotor localization. Euklidian coordinates, relative signal change and correlation to the applied reference function were determined. Statistical means were calculated. IM recruited somatotopically organized primary sensorimotor representations of the precentral gyrus that reflected the homunculus and overlapped in part with EM representations. Mean BOLD-signals ranged from 1.93 to 3.18% for EM, and from 0.73 to 1.47% for IM. The results support the hypothesis that the primary sensorimotor cortex is active during IM and that IM and EM share common functional circuits.