A Model to Identify Specific Predictors of Spatial Neglect Recovery

Overview

This study examines methods to better predict improvement of a hidden disability of functional vision, spatial neglect, following stroke. Spatial neglect is a tendency to make visual judgment and movement errors mislocating the body and objects in space. The investigators are using specialized statistical methods to compute the proportion of improvement accounted for by personal characteristics of each stroke survivor, the proportion of improvement accounted for by the unique visual-spatial errors made by each subject, and the proportion of improvement accounted for by each treatment administered. The investigators will also examine whether brain imaging predicts how rapidly improvement occurs. Lastly, the study tests whether improvements that are meaningful to the survivor can be measured in a way that still allows detection of small and scientifically eloquent performance changes.

DESCRIPTION: Spatial neglect, pathologically asymmetric spatial behavior resulting from a brain injury (Heilman, 1979) and causing functional disability (Barrett and Burkholder, 2006) may occur in 20-50% of strokes, up to 350,000 Americans annually (Ringman et al., 2004; American Stroke Assoc., 2007). Difficulty eating, dressing, and navigating in complex environments occurs acutely in this disorder, but even if symptoms improve in chronic recovery, people with spatial neglect are more likely to lose functional independence (Katz et al., 1999). Current standard clinical approaches are not theory driven, and widely-employed therapies may be only marginally effective. In this proposal, we suggest two means by which scientific acute spatial neglect treatment can be implemented. In available studies, subject heterogeneity may have obscured treatment effects. Across methods, studies used single subject, case series, and group analytic designs, but did not attempt to reconcile the distinct advantages offered by individual versus group analytic approaches. Different treatments might affect different spatial cognitive recovery functions, but simple, global outcome measures may not reflect these changes. Modeling both subject-specific and group effects is also an extremely useful method of examining targeted treatment effects. We will collect spatial neglect treatment response data over four years, for two promising and feasible spatial neglect treatments: prism adaptation training and dopaminergic medication. With mechanism-specific outcome assessment and hierarchical linear modeling, we will examine whether treatments result in predictable response. We will also examine whether controlling for subject-specific predictors models group recovery trajectory. Lastly, we will examine current standard global outcome measures instruments which have not been fully psychometrically developed, and attempt to predict subject- and group-specific recovery profiles for these variables. We hope this research will improve our ability to design cognitive rehabilitation treatment studies. It may also, however, improve our ability to translate cognitive neuroscience models of action, spatial knowledge, and attention, to treatments to optimize adaptive movement in complex environments. PUBLIC HEALTH RELEVANCE: This study investigates novel methods of outcome analysis for comparing two treatments for hidden disabilities in functional vision after stroke. We hope this research will improve our ability to design cognitive rehabilitation treatment studies. It may also, however, improve our ability to bring basic brain science to the bedside, to optimize stroke survivors' adaptive movement and balanced visual-spatial function in complex environments. Although subjects are assigned to treatments in this study, we classified the study as observational because there are no quasi-experimental options fitting its hierarchical design under the interventional study description. At no point do we examine the independent effect of treatments; rather, we assign subjects to treatments in order to learn if treatment exerts an additional effect IN COMBINATION WITH 1) individual differences and 2) performance characterization of spatial bias type immediately after stroke. We are lastly making separate examination of the predictive effect of brain lesion location in combination with variables 1) and 2). NO traditional efficacy comparisons between the two treatments to which subjects are assigned, are made. Neither treatment is considered "control" or "placebo," and there is no attempt to match severity or other relevant variables between the two treatment assignments.