Interactive Multimodal Ambulatory Monitoring to Investigate the Association between Physical Activity and Affect.

Although there is a wealth of evidence that physical activity has positive effects on psychological health, a large proportion of people are inactive. Data regarding counts, steps, and movement patterns are limited in their ability to explain why people remain inactive. We propose that multimodal ambulatory monitoring, which combines the assessment of physical activity with the assessment of psychological variables, helps to elucidate real world physical activity. Whereas physical activity can be monitored continuously, psychological variables can only be assessed at discrete intervals, such as every hour. Moreover, the assessment of psychological variables must be linked to the activity of interest. For example, if an inactive and overweight person is physically active once a week, psychological variables should be assessed during this episode. Linking the assessment of psychological variables to episodes of an activity of interest can be achieved with interactive monitoring. The primary aim of our interactive multimodal ambulatory monitoring approach was to intentionally increase the number of e-diary assessments during "active" episodes. We developed and tested an interactive monitoring algorithm that continuously monitors physical activity in everyday life. When predefined thresholds are surpassed, the algorithm triggers a signal for participants to answer questions in their electronic diary. Using data from 70 participants wearing an accelerative device for 24 h each, we found that our algorithm quadrupled the frequency of e-diary assessments during the activity episodes of interest compared to random sampling. Multimodal interactive ambulatory monitoring appears to be a promising approach to enhancing our understanding of real world physical activity and movement.

Detection of the erect position in the freely-moving human: sensor characteristics, reliability, and validity.

The present report systematically examined a means to electronically detect the erect position in the human in the natural setting. The detector was based on pressure changes in a glycerin filled tube attached to a subject's leg, and it unobtrusively and continuously measured the relative vertical distance between the hip and leg. Initial experiments established the reliability of the sensor system as a function of 1. different sizes of the tubing, 2. different amounts of air in the glycerin and 3. different ambient temperatures (6 degrees-32 degrees C). Then, in a laboratory study of normal adults, the detector was seen to discriminate sitting from standing and (when activity data were included) these two behaviors, in turn, from walking. The detector also accounted for significant differences in HR seen in the standing, as opposed to the sitting, position. In addition, when subjects carried the detector during their daily activities and provided information about their activities using an experience sampling procedure, sitting was discriminated from standing and walking with acceptable diagnostic characteristics. Thus, sitting was detected with a sensitivity of 86.1% (correct detection of all occasions when sitting actually occurred). a positive predictive value of 92.6% (occasions that the detector was right when it indicated sitting), and a negative predictive value of 80.7% (occasions that the detector was right when it indicated sitting). Finally, we demonstrated in two additional ways the direct benefit of our detector in behavioral studies in the natural environment. First, with the detector, we could confirm that a subject had performed simple activities and errands while not under close supervision. Second, cigarette smoking in the natural environment was shown to increase HR, but only when the subjects were sitting. It was concluded that our detector can be effectively applied to the identification of the sitting vs. the erect position in humans in the natural setting, and that this information may be necessary to interpret behavioral and physiological effects seen in such subjects.

Smoking produces a smaller increase in heart rate in the natural smoking environment than in the laboratory.

We tested whether the physiological effects of smoking a cigarette under standard conditions in a laboratory are similar to those seen in a room at home. On two separate test days ten healthy smokers were prepared with a small physiological recorder and were then requested to carry out a protocol requiring them to smoke one of their cigarettes in their usual way either (1) in the laboratory where they had never smoked previously, or (2) at home, alone in a quiet room where they regularly smoke. The experiment which counterbalanced the order of testing showed that in both test situations smoking produced a clear increase in heart rate (HR) and in skin conductance (SC); however, when testing was carried out in the home environment the increase in the HR was significantly less than in the laboratory. No significant differences were found for the baseline HR values in the two environments and there were no significant situational effects in the SC data. It was concluded that under conditions of normal smoking, data on the physiological effects of a cigarette in the laboratory may not be fully generalizable to those seen in the natural smoking situation. Several possible mechanisms of this situationally-specific effect of smoking were discussed.