Study Timeline and Data Collection

Data will be collected from intervention and control group participants at initial assessment (T0), prior to delivery of 3MDR (T1), pre/post 3MDR sessions (T2.1-2.6), and at 1 week (T3), 1 month (T4), 3 months (T5), and 6 months (T6) postintervention (Figure 3).

Multimodular motion-assisted memory desensitization and reconsolidation (3MDR) data collection time points for the intervention (treatment; blue/red) and waitlist control (green) groups. The syringe icon denotes blood collection and the square icon denotes saliva collection.

Clinical outcome measures will be collected at each time point using the questionnaires listed in Table 1. Qualitative interviews will occur at T3 through T6. Physiological, biometric, and qualitative audio/video data will be collected at T2.1-T2.6. Blood and saliva samples will be collected at baseline (T1), before (pre) and immediately after (post) 3MDR sessions at T2.1 and T2.6, post 3MDR sessions at T2.3 and T2.4, at 1-week follow up (T3), and at 1-, 3-, and 6-month follow ups (T4, T5, T6) for the analyses of peripheral blood and salivary biomarkers, including inflammatory mediators, neuroendocrine hormones, and microRNAs (see Multimedia Appendix 1).

Demographic information will include gender, age, ethnicity, marital status, military branch (ie, army, navy, or air force), and occupation. Psychological outcomes will be assessed through paper-based questionnaires administered during each contact with participants (Table 1). For further information on each outcome measure, refer to Multimedia Appendix 2.

Walking and eye-scanning patterns, as well as physiological data (eg, heart rate, breathing rate, gait pattern, force plate analysis) will be collected during the training session through the CAREN system, Tobii Mobile eye-tracking glasses, Muse electroencephalogram (EEG), and Zephyr BioHarness 3.

Video recording of each session will capture qualitative data of the exchange between the clinician and participant during the 3MDR session. Audio recordings of the therapeutic debriefs will also occur, during which the clinician and patient will discuss the experience of any remarkable or meaningful aspects of the session. The clinician will also share observations. Upon completion of the 3MDR intervention, audio-recorded iterative semistructured interviews will be conducted with all participants at week 10, and 1, 3, and 6 months postintervention to explore their experience of 3MDR, its impact on their PTSD symptoms and overall function, and participation in the research study.

To measure the perceived technology acceptance and usability of 3MDR and the CAREN, two 15-question surveys based on the Unified Theory of Acceptance and Usability of Technology (UTAUT) model [56] will be administered pre/post exposure to CAREN and 3MDR. The surveys utilized Likert scores ranging from 15 to 105 points, with high scores indicating increased perceived technology acceptance and usability.

The impact of delivering 3MDR will be examined with clinicians and operators using questionnaires related to technological acceptance, perceived stress, and professional quality of life prior to 3MDR training and on completion of a full course of 3MDR with a study participant (Table 2).

Time points of data collection for multimodular motion-assisted memory desensitization and reconsolidation (3MDR) clinicians and operators.

^aUTAUT: Unified Theory of Acceptance and Usability of Technology.

^bPSS: Perceived Stress Scale.

^cSUDS: Subjective Units of Distress Scale.

^dPQoL: Professional Quality of Life Scale.

^eSTSS: Secondary Traumatic Stress Scale.

Although group-based statistical analysis is useful for drawing conclusions about general trends, it has limited utility for guiding individual diagnostic and treatment choices.

As a result, in addition to conventional statistical analysis, this study will explore the application of machine learning in an attempt to identify individual data patterns that may predict diagnosis, severity, and potential treatment outcomes. Physiological, biometric, biological, and qualitative data from T0 to T3 will be used to predict psychological outcome measures at week 9 plus follow-up sessions. A model that can compute sequential, multimodal data is needed to autonomously observe and learn the variability of the relevant feature sets and its association with outcome measures [57,58]. Deep neural networks will be applied for this task [57,58].