Journal of Rehabilitation Medicine 51-9 | Page 25

D: –0.021 to 0.019 –0.028 to 0.014 –0.005 –0.007 –0.002 0.316 –0.042 to 0.004 p-value Physical activity post-stroke to prevent cognitive impairment 649 ment in cognitive function may follow a similar pattern (20), suggesting that earlier initiation, and perhaps more specific and intensive, future interventions would be beneficial. In addition, the intervention was based on patient’s preferences more than the type of training aiming at cognitive gain. Combining aerobic exercise with other training programmes seems to achieve bet- ter cognitive function and might be applied in future trials (7, 21). The lack of effect on cognition might be explained by lack of adherence to the intervention. However, another possible explanation for no effect on cogni- tion is the heterogeneity of the stroke population and, probably also, the underlying brain pathology (22). Theoretically, the effect of physical activity on brain function includes neurogenesis, building brain reserve, angiogenesis and prevention of progressive small ves- sel disease (23). Light-to-moderate physical activity may represent a treatment strategy less likely to affect the underlying pathology in the chronic phase. How­ ever, it might halt or slow decline through building brain reserve and thereby delay symptom onset. A po- tential benefit on vascular pathology, to prevent cogni- tive dysfunction due to chronic cerebral hypoperfusion may need longer follow-up. Physical activity could have a beneficial effect on cognitive flexibility and brain reserve, but an effect through neurogenesis and angiogenesis may be achieved through interventions with higher intensities (22). Secondary prevention after cerebrovascular disease involves both pharmacological treatment and interventions for behaviourally modi- fiable risk factors (24). A multimodal individualized intervention including optimal medication, dietary, social and cognitive stimulation in addition to physical activity might also be needed to improve cognition after stroke, perhaps with longer follow-up (25–27). Post-stroke, there is also a wide spectrum of neuropsychological manifestations. Apparent brain damage gives rise to different cognitive deficits, es- pecially executive function, attention and processing speed and working memory (28, 29). It is a limitation that HADS, MMSE and the Trail Making Tests are primarily designed as screening tools and might not be sensitive to detect changes that occur during follow-up after stroke. Hence, future research should investigate the psychometric properties of these measures in the stroke population. Furthermore, the study population may, due to the rather high mean age, have been too heterogeneous regarding both underlying additional degeneration and comorbidities. In addition, the study was designed or powered for a potential effect on functional outcome, and length of education was not registered. Furthermore, functional impairments in the dominant arm/hand may have affected the results J Rehabil Med 51, 2019