Heart rate variability (HRV) rapidly gains attention as an important marker of cardiovascular autonomic modulation. Moreover, there is evidence for a link between the autonomic deficit measurable by reduced HRV and the hypoactivity of the cholinergic system, which is prominently affected in Alzheimer's disease (AD). Despite the positive influence of physical exercise on cognition and its promising association with HRV, previous studies did not explore the effect of long-term physical exercise in older adults with AD. Taking advantage of a longitudinal study we analyzed the effect of a 20-week dual task training regime (3 × 15-min per week) on the vagal mediated HRV index RMSSD (root mean square of successive RR interval differences) during physical exercise and the short-term memory performance in a AD cohort (N = 14). Each training contained physical exercise on a bicycle ergometer while memorizing 30 successively presented pictures as well as the associated post-exercise picture recognition memory test. Linear-mixed modeling revealed that HRV-RMSSD significantly increased over the intervention time. Moreover, the reaction time in the picture recognition task decreased while the accuracy remained stable. Furthermore, a significantly negative relationship between increased fitness measured by HRV-RMSSD and decreased reaction time was observed. This feasibility study points to the positive effects of a dual task regime on physical and cognitive fitness in a sample with impaired cognitive performance. Beyond this, the results show that the responsiveness of parasympathetic system as measured with HRV can be improved in patients with dementia.

Various longitudinal studies report positive effects of physical exercise on cognitive performance in patients with Alzheimer's disease (AD). However, most of these studies fail to capture the dynamic nature of the change and interplay of physical fitness and cognition. To overcome this limitation, we employ a hierarchical continuous-time dynamic model to analyze data of a 24-week physical and cognitive dual-task training regime (3 sessions per week) in older adults with AD (N = 17). The model was specified with two fully connected state variables enabling bidirectional effects between physical and cognitive performance over measurement occasions. The results show that physical performance is dynamically linked to cognitive performance. In line with earlier findings, higher physical performance predicted improved memory recognition performance in terms of faster reaction times.