Sampling frequency impacts measurement of walking activity after stroke

Area of Science:

• Rehabilitation Science
• Biomedical Engineering
• Movement Science

Background:

• Understanding daily physical activity is crucial for stroke recovery.
• Accurate measurement of walking patterns informs rehabilitation strategies.
• Previous research has not fully explored the impact of data sampling on activity metrics.

Purpose of the Study:

• To investigate how different sampling epochs affect the measurement of total walking time and the number of walking bouts per day in individuals post-stroke.
• To determine the sensitivity of these activity variables to sampling epoch duration, particularly during periods of low and high daily activity.

Main Methods:

• Ninety-eight stroke survivors (average age 63.8 years) participated.
• Activity was monitored using StepWatch Activity Monitors for 3-5 days.
• Stride data collected in 5-second epochs were down-sampled to 10, 20, 30, and 60-second epochs.

Main Results:

• Total walking time and number of walking bouts varied significantly across all sampling epochs (p < 0.001).
• A 5-second epoch captured approximately 40% more walking time and 6 times more bouts than a 60-second epoch.
• These differences were more pronounced on low-activity days.

Conclusions:

• The choice of sampling epoch critically influences the assessment of daily walking activity in stroke patients.
• Researchers must carefully select sampling epochs to ensure accurate and reliable data, especially when analyzing low-activity periods.
• Consideration of sampling epoch is essential for designing robust studies on post-stroke mobility and rehabilitation outcomes.