A Bayesian Assessment of Real-World Behavior During Multitasking.

Area of Science:

• Biomechanics
• Human-Computer Interaction
• Cognitive Science

Background:

• Multitasking is prevalent in daily life, yet its impact on activities of daily living (ADLs) remains unclear.
• Assessing motor performance during ADLs in both healthy individuals and patient populations is crucial for understanding functional capacity.

Purpose of the Study:

• To investigate the effects of multitasking on motor performance during a common ADL: meal preparation.
• To quantify changes in motor activity patterns under single-task versus multitask conditions.

Main Methods:

• Healthy participants performed meal preparation tasks under single and multiple cognitive load conditions.
• A wearable sensor network monitored motor activities, quantifying hand trajectories and wrist accelerations.
• A Naïve Bayes Model estimated the probability of multitasking based on motor data, distinguishing between single and triple task conditions.

Main Results:

• Wrist accelerometer data demonstrated distinct task patterns, particularly in high and low value ranges.
• The probability of specific motor performance patterns changed significantly when participants engaged in additional cognitive tasks.
• In healthy individuals, multitasking increased the likelihood of observing lower wrist acceleration frequency patterns.

Conclusions:

• Multitasking demonstrably alters motor performance during everyday activities like meal preparation.
• Changes in motor patterns, specifically reduced acceleration frequencies, are indicative of increased cognitive load.
• These findings suggest that cognitive decline associated with aging or disease may exacerbate the motor performance differences observed during multitasking.