Combined Cognitive and Sensorimotor Assessment Enhances Detection of Cognitive Declines

Introduction: Alzheimer’s Disease and related dementias (ADRD) impact an estimated 6.7 million adults in the United States, with projections indicating a rise to 13.8 million by 2060 (Alzheimer’s Association, 2023). ADRD represents a progressive neurodegenerative condition currently devoid of a cure. Optimal mitigation of ADRD progression and the preservation of cognitive function are most achievable when interventions are administered early in the disease process. However, current clinical cognitive assessments face challenges in sensitively characterizing various forms of mild cognitive impairment (MCI) and ADRD due to their inability to discern subtle cognitive differences. As expounded in NOT-AG-22-044 and associated literature, cognitive and sensory-motor alterations are intricately linked. Consequently, human sensorimotor changes hold means as potential biomarkers for MCI/ADRD progression. Although recent research has demonstrated the capacity of saccadic eye movements to detect subtle cognitive declines, the integration of multiple assessment modalities through multimodal and multivariate approaches may yield more comprehensive insights, enhancing MCI/ADRD diagnosis and classification.
Hence, there exists a demand for sensitive multimodal assessments capable of identifying subtle cognitive alterations throughout the progression of MCI and ADRD. This study aims to: 1) devise a multimodal quantitative assessment technique by integrating eye-tracking cameras and an arm-hand wearable robot, 2) develop combined cognitive and sensorimotor tests based on widely-used cognitive assessments, namely the trail-making test B (TMT-B) and anti-saccade test (AST), and 3) conduct combined simultaneous cognitive and sensory/motor tests to evaluate the performance of older and young adult groups.

Materials and

Methods: A novel system was developed by integrating a multi-joint arm-hand robot and a head-mounted eye-tracking camera (Fig. 1). The robot joint movement was monitored in real-time to analyze subjects’ sensory/motor performance; the eye-tracking camera captured saccadic data at 500Hz, enabling quantitative cognitive assessment. The AST was used during which subjects were asked to fixate on a motionless target and make a saccade in the direction opposite to the stimulus that was presented to one side of the target. Three combined tests were created, including (1) Combined TMT-B (CTMT-B): subjects looked at a series of numbers (1 to 13) and letters (A to L) in ascending order and alternated between numbers and letters while moving the cursor on the screen to the numbers or letters they focused on simultaneously through the exoskeleton, (2) Combined AST (CAST): the subjects performed AST with the dominant arm wearing the robot which fixes the arm-hand at a predefined posture, and flexed/extended their wrist joints to the direction they looked to simultaneously, and (3) Combined Sensory test (CST): the robot slowly flexed/extended the wrist joint, and when subjects felt the wrist joint motion, they were required to immediately look to the opposite direction to the robot's motion and flex/extend the wrist joints to the same direction simultaneously.
Ten healthy young (five males and five females, age: 27±3.7) and eleven healthy older (five males and six females, age: 65±7.7) subjects were recruited. Upon obtaining written informed consent, participants completed AST, CAST, CST, and CTMT-B.

Results, Conclusions, and Discussions: Participants’ correct eye movement latency time was employed to evaluate their cognitive performance during AST and CAST. As shown in Fig. 2, paired t-test identified that the older adults showed significantly increased correct eye movement latency time during CAST compared with AST (p=0.013, t10=3.00), whereas no significant changes in young subjects were observed (p=0.501, t9=0.70). Therefore, the proposed CAST can potentially improve the detection of cognitive changes. The subjects’ correct hand movement latency time was utilized to quantify their motor performance during CAST. The two-sample t-test found that the older adults exhibited both significantly longer eye (p=0.001, t19=3.82) and hand (p=0.037, t19=2.24) movement latency time, which indicates that the older adults showed weaker cognitive and motor function than the younger adults. For CST, the two-sample t-test revealed significantly increased eye movement latency time of the older subjects than the young ones (p=0.028, t19=2.37), as shown in Fig. 3. The older adults also showed marginally longer wrist movement latency time than the young adults (p=0.063, t19=1.98).
For CTMT-B, the total time to complete the test, the mean reach time (average time the cursor moved between targets), and mean dwell time (average time the cursor was located on targets) were calculated to evaluate the overall performance. Moreover, the total number of saccades during the test was utilized to evaluate participants’ cognitive function; the cursor movement smoothness (number of cursor motion speed peaks during reaching movements) was used to assess subjects’ motor function. As summarized in Fig. 4, the older subjects exhibited significantly increased total time (p < 0.001, t18=4.66), average dwell (p=0.011, t18=2.81), and reach time (p=0.001, t18=3.92), compared to the young subjects. For the cursor movement smoothness measurement, the older adults showed a significantly larger number of velocity peaks (p < 0.001, t18=5.92). Regarding the number of saccades, although no significant differences were observed (p=0.083, t16=1.85), older adults performed marginally increased value than the young subjects (Fig. 5).
In summary, the results support that the combined cognitive and sensorimotor tests can quantify human cognitive and motor functions more accurately and promote the identification of cognitive declines.

Acknowledgements (Optional):