Accurate outcome prediction following stroke is important for proper delivery of poststroke care. It has been difficult to determine specific factors that provide reliable and accurate predictions of outcome, particularly for patients with intermediate deficit severities. Age and severity of deficit have repeatedly been found to be most reliable, but only as rough estimates and for patients at either extreme of the disability spectrum. This paper reports a prospective study of consecutive rehabilitation admissions (N = 536) to determine the influence of preselected factors. Outcome was analyzed in terms of functional improvement and disposition. Patients younger than 55 years or with an admission Functional Independence Measure (FIM) greater than 80 almost universally went home. Admission FIMs less than 40 were associated with nearly certain nursing home discharge. The comprehensive FIM score was a stronger predictor of outcome than motor impairment in isolation. An admission FIM of 60 or greater was associated with a higher probability of functional improvement during rehabilitation. Small-vessel strokes had the best outcome. Intracerebral hemorrhages improved more than ischemic strokes but more slowly. Right hemisphere lesions did worse than left. Comorbidities influenced outcome only when several conditions accumulated. The absence of a committed caregiver at home increased the risk of nursing home discharge. Suggestions for rehabilitation triage are given.

Posterior parietal cortex is active during episodic memory retrieval, yet its role remains unclear. According to a recent proposal, dorsal parietal cortex (DPC) allocates top-down attention to memory retrieval, whereas ventral parietal cortex (VPC) mediates the bottom-up attentional capture by retrieved contents, i.e., the Attention-to-Memory (AtoM) hypothesis. Here, for the first time, functional magnetic resonance imaging (fMRI) and lesion techniques were combined in a single study to test the role of parietal cortex in episodic retrieval. Participants studied word pairs and then detected studied (target) words among new words. In some conditions, a studied word cued the upcoming target word, facilitating recognition performance. In line with the AtoM hypothesis, left DPC was engaged when participants searched for/anticipated memory targets upon presentation of relevant memory cues and predicted the ensuing behavioral advantage. In contrast, left VPC predicted efficacy and speed of target detection on noncued trials and was largest for memory targets that were invalidly cued. Consistent with fMRI evidence, patients with lesions in DPC did not benefit from memory cueing, whereas patients with lesions in VPC had problems recognizing unexpected memory targets. These results support the AtoM hypothesis that DPC and VPC mediate top-down and bottom-up attention to memory retrieval, respectively.

Background and Purpose— Many ischemic strokes or transient ischemic attacks are labeled cryptogenic but may have undetected atrial fibrillation (AF). We sought to identify those most likely to have subclinical AF. Methods— We prospectively studied patients with cryptogenic stroke or transient ischemic attack aged ≥55 years in sinus rhythm, without known AF, enrolled in the intervention arm of the 30 Day Event Monitoring Belt for Recording Atrial Fibrillation After a Cerebral Ischemic Event (EMBRACE) trial. Participants underwent baseline 24-hour Holter ECG poststroke; if AF was not detected, they were randomly assigned to 30-day ECG monitoring with an AF auto-detect external loop recorder. Multivariable logistic regression assessed the association between baseline variables (Holter-detected atrial premature beats [APBs], runs of atrial tachycardia, age, and left atrial enlargement) and subsequent AF detection. Results— Among 237 participants, the median baseline Holter APB count/24 h was 629 (interquartile range, 142–1973) among those who subsequently had AF detected versus 45 (interquartile range, 14–250) in those without AF ( P <0.001). APB count was the only significant predictor of AF detection by 30-day ECG ( P <0.0001), and at 90 days ( P =0.0017) and 2 years ( P =0.0027). Compared with the 16% overall 90-day AF detection rate, the probability of AF increased from <9% among patients with <100 APBs/24 h to 9% to 24% in those with 100 to 499 APBs/24 h, 25% to 37% with 500 to 999 APBs/24 h, 37% to 40% with 1000 to 1499 APBs/24 h, and 40% beyond 1500 APBs/24 h. Conclusions— Among older cryptogenic stroke or transient ischemic attack patients, the number of APBs on a routine 24-hour Holter ECG was a strong dose-dependent independent predictor of prevalent subclinical AF. Those with frequent APBs have a high probability of AF and represent ideal candidates for prolonged ECG monitoring for AF detection. Clinical Trial Registration— URL: http://www.clinicaltrials.gov . Unique identifier: NCT00846924.

Subject-specific artifacts caused by head motion and physiological noise are major confounds in BOLD fMRI analyses. However, there is little consensus on the optimal choice of data preprocessing steps to minimize these effects. To evaluate the effects of various preprocessing strategies, we present a framework which comprises a combination of (1) nonparametric testing including reproducibility and prediction metrics of the data-driven NPAIRS framework (Strother et al. [2002]: NeuroImage 15:747-771), and (2) intersubject comparison of SPM effects, using DISTATIS (a three-way version of metric multidimensional scaling (Abdi et al. [2009]: NeuroImage 45:89-95). It is shown that the quality of brain activation maps may be significantly limited by sub-optimal choices of data preprocessing steps (or "pipeline") in a clinical task-design, an fMRI adaptation of the widely used Trail-Making Test. The relative importance of motion correction, physiological noise correction, motion parameter regression, and temporal detrending were examined for fMRI data acquired in young, healthy adults. Analysis performance and the quality of activation maps were evaluated based on Penalized Discriminant Analysis (PDA). The relative importance of different preprocessing steps was assessed by (1) a nonparametric Friedman rank test for fixed sets of preprocessing steps, applied to all subjects; and (2) evaluating pipelines chosen specifically for each subject. Results demonstrate that preprocessing choices have significant, but subject-dependant effects, and that individually-optimized pipelines may significantly improve the reproducibility of fMRI results over fixed pipelines. This was CIHR Author Manuscript CIHR Author Manuscript CIHR Author Manuscriptdemonstrated by the detection of a significant interaction with motion parameter regression and physiological noise correction, even though the range of subject head motion was small across the group (≪ 1 voxel). Optimizing pipelines on an individual-subject basis also revealed brain activation patterns either weak or absent under fixed pipelines, which has implications for the overall interpretation of fMRI data, and the relative importance of preprocessing methods.