Methods for quantifying change in ambulatory behavious

INTRODUCTION Objective measures of ambulation are increasingly used to assess the impact of public health or rehabilitation intervention. These instruments enable long-term, free-living recording of walking behaviour. This offers a unique opportunity to assess changes in behaviour. However analysis of objective data is often restricted to quantifying the volume of walking (e.g. time spent walking, number of steps, distance), which due to high intra and inter subject variability do not enable to quantify change in behaviour with much sensitivity. We have developed methods that quantify the diversity, complexity and variability of walking behaviour that can be used to assess subtle changes in behaviour. METHODS 17 participants with advanced Parkinson Disease (UPDRS III 29.1+5.5; disease duration 12.5+6.4; HY (3 4) took part in this longitudinal study. Ambulatory activity was measured objectively over a 7 day period using an activPALTM. Data were collected 6 weeks before surgery and 6 weeks, 3, 6 and 12 months after surgery. We examine the distribution of length and cadences of walking bouts in free-living conditions and how fragmented walking bouts are 2. Based on these statistics we compute the probability P(l,c) to find a walking bouts of length l and cadence c. Using Hills numbers Da= exp( 1 we derive measures of the diversity, complexity and variability of walking bouts used by an individual. These numbers are normalised and expressed as al number of walking bouts. RESULTS The distribution of walking bouts follows a log-normal distribution and can be characterised standard log-normal statistics. Cadence is distributed normally. The possible combination of length and cadence pairs can therefore be represented by partition of the log of bout length and linear cadences centered around the mean cadence. Mixed model analysis shows that volume of walking is unaffected by brain stimulation surgery. However the walking behaviour has changed with more less fragmented more variable and diverse ambulatory periods. DISCUSSION AND CONCLUSION The results hint that showing that brain stimulation enabled Parkinson Disease patients to regain more flexibility and adaptability in their walking, allowing for longer more diverse ambulatory periods. Methods developed to quantify ambulatory behaviour have the potential to inform intervention and their outcomes. Diversity of ambulation has also been shown to be an important factor to be considered along side the volume of walking in the assessment of physical activity in older adult3.