Better understanding fall risk: AI-based computer vision for contextual gait assessment.

Contemporary research to better understand free-living fall risk assessment in Parkinson's disease (PD) often relies on the use of wearable inertial-based measurement units (IMUs) to quantify useful temporal and spatial gait characteristics (e.g., step time, step length). Although use of IMUs is useful to understand some intrinsic PD fall-risk factors, their use alone is limited as they do not provide information on extrinsic factors (e.g., obstacles). Here, we update on the use of ergonomic wearable video-based eye-tracking glasses coupled with AI-based computer vision methodologies to provide information efficiently and ethically in free-living home-based environments to better understand IMU-based data in a small group of people with PD. The use of video and AI within PD research can be seen as an evolutionary step to improve methods to understand fall risk more comprehensively.

Using Video Technology and AI within Parkinson's Disease Free-Living Fall Risk Assessment.

Falls are a major concern for people with Parkinson's disease (PwPD), but accurately assessing real-world fall risk beyond the clinic is challenging. Contemporary technologies could enable the capture of objective and high-resolution data to better inform fall risk through measurement of everyday factors (e.g., obstacles) that contribute to falls. Wearable inertial measurement units (IMUs) capture objective high-resolution walking/gait data in all environments but are limited by not providing absolute clarity on contextual information (i.e., obstacles) that could greatly influence how gait is interpreted. Video-based data could compliment IMU-based data for a comprehensive free-living fall risk assessment. The objective of this study was twofold. First, pilot work was conducted to propose a novel artificial intelligence (AI) algorithm for use with wearable video-based eye-tracking glasses to compliment IMU gait data in order to better inform free-living fall risk in PwPD. The suggested approach (based on a fine-tuned You Only Look Once version 8 (YOLOv8) object detection algorithm) can accurately detect and contextualize objects (mAP50 = 0.81) in the environment while also providing insights into where the PwPD is looking, which could better inform fall risk. Second, we investigated the perceptions of PwPD via a focus group discussion regarding the adoption of video technologies and AI during their everyday lives to better inform their own fall risk. This second aspect of the study is important as, traditionally, there may be clinical and patient apprehension due to ethical and privacy concerns on the use of wearable cameras to capture real-world video. Thematic content analysis was used to analyse transcripts and develop core themes and categories. Here, PwPD agreed on ergonomically designed wearable video-based glasses as an optimal mode of video data capture, ensuring discreteness and negating any public stigma on the use of research-style equipment. PwPD also emphasized the need for control in AI-assisted data processing to uphold privacy, which could overcome concerns with the adoption of video to better inform IMU-based gait and free-living fall risk. Contemporary technologies (wearable video glasses and AI) can provide a holistic approach to fall risk that PwPD recognise as helpful and safe to use.

Assessment of Visual Attention in Teams with or without Dedicated Team Leaders: A Neonatal Simulation-Based Pilot Randomised Cross-Over Trial Utilising Low-Cost Eye-Tracking Technology.

BACKGROUND: Eye-tracking technology could be used to study human factors during teamwork. OBJECTIVES: This work aimed to compare the visual attention (VA) of a team member acting as both a team leader and managing the airway, compared to a team member performing the focused task of managing the airway in the presence of a dedicated team leader. This work also aimed to report differences in team performance, behavioural skills, and workload between the two groups using validated tools. METHODS: We conducted a simulation-based, pilot randomised controlled study. The participants included were volunteer paediatric trainees, nurse practitioners, and neonatal nurses. Three teams consisting of four team members were formed. Each team participated in two identical neonatal resuscitation simulation scenarios in a random order, once with and once without a team leader. Using a commercially available eye-tracking device, we analysed VA regarding attention to (1) a manikin, (2) a colleague, and (3) a monitor. Only the trainee who was the airway operator would wear eye-tracking glasses in both simulations. RESULTS: In total, 6 simulation scenarios and 24 individual role allocations were analysed. Participants in a no-team-leader capacity had a greater number of total fixations on manikin and monitors, though this was not significant. There were no significant differences in team performance, behavioural skills, and individual workload. Physical demand was reported as significantly higher by participants in the group without a team leader. During debriefing, all the teams expressed their preference for having a dedicated team leader. CONCLUSION: In our pilot study using low-cost technology, we could not demonstrate the difference in VA with the presence of a team leader.

iVOMS: Instrumented Vestibular / Ocular motor screen in healthy controls and mild traumatic brain injury.

Objective Vestibular/ocular deficits occur with mild traumatic brain injury (mTBI). The vestibular/ocular motor screening (VOMS) tool is used to assess individuals post-mTBI, which primarily relies upon subjective self-reported symptoms. Instrumenting the VOMS (iVOMS) with technology may allow for more objective assessment post-mTBI, which reflects actual task performance. This study aimed to validate the iVOMS analytically and clinically in mTBI and controls. Methods Seventy-nine people with sub-acute mTBI (<12 weeks post-injury) and forty-four healthy control participants performed the VOMS whilst wearing a mobile eye-tracking on a one-off visit. People with mTBI were included if they were within 12 weeks of a physician diagnosis. Participants were excluded if they had any musculoskeletal, neurological or sensory deficits which could explain dysfunction. A series of custom-made eye tracking algorithms were used to assess recorded eye-movements. Results The iVOMS was analytically valid compared to the reference (ICC2,1 0.85-0.99) in mTBI and controls. The iVOMS outcomes were clinically valid as there were significant differences between groups for convergence, vertical saccades, smooth pursuit, vestibular ocular reflex and visual motion sensitivity outcomes. However, there was no significant relationship between iVOMS outcomes and self-reported symptoms. Conclusion The iVOMS is analytically and clinically valid in mTBI and controls, but further work is required to examine the sensitivity of iVOMS outcomes across the mTBI spectrum. Findings also highlighted that symptom and physiological issue resolution post-mTBI may not coincide and relationships need further examination.

Enhancing fall risk assessment: instrumenting vision with deep learning during walks.

BACKGROUND: Falls are common in a range of clinical cohorts, where routine risk assessment often comprises subjective visual observation only. Typically, observational assessment involves evaluation of an individual's gait during scripted walking protocols within a lab to identify deficits that potentially increase fall risk, but subtle deficits may not be (readily) observable. Therefore, objective approaches (e.g., inertial measurement units, IMUs) are useful for quantifying high resolution gait characteristics, enabling more informed fall risk assessment by capturing subtle deficits. However, IMU-based gait instrumentation alone is limited, failing to consider participant behaviour and details within the environment (e.g., obstacles). Video-based eye-tracking glasses may provide additional insight to fall risk, clarifying how people traverse environments based on head and eye movements. Recording head and eye movements can provide insights into how the allocation of visual attention to environmental stimuli influences successful navigation around obstacles. Yet, manual review of video data to evaluate head and eye movements is time-consuming and subjective. An automated approach is needed but none currently exists. This paper proposes a deep learning-based object detection algorithm (VARFA) to instrument vision and video data during walks, complementing instrumented gait. METHOD: The approach automatically labels video data captured in a gait lab to assess visual attention and details of the environment. The proposed algorithm uses a YoloV8 model trained on with a novel lab-based dataset. RESULTS: VARFA achieved excellent evaluation metrics (0.93 mAP50), identifying, and localizing static objects (e.g., obstacles in the walking path) with an average accuracy of 93%. Similarly, a U-NET based track/path segmentation model achieved good metrics (IoU 0.82), suggesting that the predicted tracks (i.e., walking paths) align closely with the actual track, with an overlap of 82%. Notably, both models achieved these metrics while processing at real-time speeds, demonstrating efficiency and effectiveness for pragmatic applications. CONCLUSION: The instrumented approach improves the efficiency and accuracy of fall risk assessment by evaluating the visual allocation of attention (i.e., information about when and where a person is attending) during navigation, improving the breadth of instrumentation in this area. Use of VARFA to instrument vision could be used to better inform fall risk assessment by providing behaviour and context data to complement instrumented e.g., IMU data during gait tasks. That may have notable (e.g., personalized) rehabilitation implications across a wide range of clinical cohorts where poor gait and increased fall risk are common.

Contextualizing remote fall risk: Video data capture and implementing ethical AI.

Wearable inertial measurement units (IMUs) are being used to quantify gait characteristics that are associated with increased fall risk, but the current limitation is the lack of contextual information that would clarify IMU data. Use of wearable video-based cameras would provide a comprehensive understanding of an individual's habitual fall risk, adding context to clarify abnormal IMU data. Generally, there is taboo when suggesting the use of wearable cameras to capture real-world video, clinical and patient apprehension due to ethical and privacy concerns. This perspective proposes that routine use of wearable cameras could be realized within digital medicine through AI-based computer vision models to obfuscate/blur/shade sensitive information while preserving helpful contextual information for a comprehensive patient assessment. Specifically, no person sees the raw video data to understand context, rather AI interprets the raw video data first to blur sensitive objects and uphold privacy. That may be more routinely achieved than one imagines as contemporary resources exist. Here, to showcase/display the potential an exemplar model is suggested via off-the-shelf methods to detect and blur sensitive objects (e.g., people) with an accuracy of 88%. Here, the benefit of the proposed approach includes a more comprehensive understanding of an individual's free-living fall risk (from free-living IMU-based gait) without compromising privacy. More generally, the video and AI approach could be used beyond fall risk to better inform habitual experiences and challenges across a range of clinical cohorts. Medicine is becoming more receptive to wearables as a helpful toolbox, camera-based devices should be plausible instruments.

Feasibility of using low-cost markerless motion capture for assessing functional outcomes after lower extremity musculoskeletal cancer surgery.

BACKGROUND: Physical limitations are frequent and debilitating after sarcoma treatment. Markerless motion capture (MMC) could measure these limitations. Historically expensive cumbersome systems have posed barriers to clinical translation. RESEARCH QUESTION: Can inexpensive MMC [using Microsoft KinectTM] assess functional outcomes after sarcoma surgery, discriminate between tumour sub-groups and agree with existing assessments? METHODS: Walking, unilateral stance and kneeling were measured in a cross-sectional study of patients with lower extremity sarcomas using MMC and standard video. Summary measures of temporal, balance, gait and movement velocity were derived. Feasibility and early indicators of validity of MMC were explored by comparing MMC measures i) between tumour sub-groups; ii) against video and iii) with established sarcoma tools [Toronto Extremity Salvage Score (TESS)), Musculoskeletal Tumour Rating System (MSTS), Quality of life-cancer survivors (QoL-CS)]. Statistical analysis was conducted using SPSS v19. Tumour sub-groups were compared using Mann-Whitney U tests, MMC was compared to existing sarcoma measures using correlations and with video using Intraclass correlation coefficient agreement. RESULTS: Thirty-four adults of mean age 43 (minimum value-maximum value 19-89) years with musculoskeletal tumours in the femur (19), pelvis/hip (3), tibia (9), or ankle/foot (3) participated; 27 had limb sparing surgery and 7 amputation. MMC was well-tolerated and feasible to deliver. MMC discriminated between surgery groups for balance (p<0.05*), agreed with video for kneeling times [ICC = 0.742; p = 0.001*] and showed moderate relationships between MSTS and gait (p = 0.022*, r = -0.416); TESS and temporal outcomes (p = 0.016* and r = -0.0557*), movement velocity (p = 0.021*, r = -0.541); QoL-CS and balance (p = 0.027*, r = 0.441) [* = statistical significance]. As MMC uncovered important relationships between outcomes, it gave an insight into how functional impairments, balance, gait, disabilities and quality of life (QoL) are associated with each other. This gives an insight into mechanisms of poor outcomes, producing clinically useful data i.e. data which can inform clinical practice and guide the delivery of targeted rehabilitation. For example, patients presenting with poor balance in various activities can be prescribed with balance rehabilitation and those with difficulty in movements or activity transitions can be managed with exercises and training to improve the quality and efficiency of the movement. SIGNIFICANCE: In this first study world-wide, investigating the use of MMC after sarcoma surgery, MMC was found to be acceptable and feasible to assess functional outcomes in this cancer population. MMC demonstrated early indicators of validity and also provided new knowledge that functional impairments are related to balance during unilateral stance and kneeling, gait and movement velocity during kneeling and these outcomes in turn are related to disabilities and QoL. This highlighted important relationships between different functional outcomes and QoL, providing valuable information for delivering personalised rehabilitation. After completing future validation work in a larger study, this approach can offer promise in clinical settings. Low-cost MMC shows promise in assessing patient's impairments in the hospitals or their homes and guiding clinical management and targeted rehabilitation based on novel MMC outcomes affected, therefore providing an opportunity for delivering personalised exercise programmes and physiotherapy care delivery for this rare cancer.

Coproduction and Usability of a Smartphone App for Falls Reporting in Parkinson Disease.

OBJECTIVE: The purpose of this study was to coproduce a smart-phone application for digital falls reporting in people with Parkinson disease (PD) and to determine usability using an explanatory mixed-methods approach. METHODS: This study was undertaken in 3 phases. Phase 1 was the development phase, in which people with PD were recruited as co-researchers to the project. The researchers, alongside a project advisory group, coproduced the app over 6 months. Phase 2 was the implementation phase, in which 15 people with PD were invited to test the usability of the app. Phase 3 was the evaluation phase, in which usability was assessed using the systems usability scale by 2 focus groups with 10 people with PD from phase 2. RESULTS: A prototype was successfully developed by researchers and the project advisory group. The usability of the app was determined as good (75.8%) by people with PD when rating using the systems usability scale. Two focus groups (n = 5 per group) identified themes of 1) usability, 2) enhancing and understanding management of falls, and 3) recommendations and future developments. CONCLUSIONS: A successful prototype of the iFall app was developed and deemed easy to use by people with PD. The iFall app has potential use as a self-management tool for people with PD alongside integration into clinical care and research studies. IMPACT: This is the first digital outcome tool to offer reporting of falls and near-miss fall events. The app may benefit people with PD by supporting self-management, aiding clinical decisions in practice, and providing an accurate and reliable outcome measure for future research. LAY SUMMARY: A smartphone application designed in collaboration with people who have PD to record their falls was acceptable and easy to use by people with PD.

Better informing everyday fall risk assessment: experimental studies with contemporary technologies.

BACKGROUND: Age-related mobility issues and frailty are a major public health concern because of an increased risk of falls. Subjective assessment of fall risk in the clinic is limited, failing to account for an individual's habitual activities in the home or community. Equally, objective mobility trackers for use in the home and community lack extrinsic (ie, environmental) data capture to comprehensively inform fall risk. We propose a contemporary approach that combines artificial intelligence (AI) and video glasses to augment current methods of fall risk assessment. METHODS: Two case studies were performed to provide a framework to assess extrinsic factors within fall risk assessment via video glasses. The first was AI-based detection of environment and terrain type. We developed convolutional neural networks (CNN) via a bespoke dataset (>145 000 images) captured from different settings (eg, offices, high streets) via free-licenced video on social media. AI automated a textual description to uphold privacy while describing the scene (eg, indoor and carpet). In the second case study, we provided video glasses to participants within a university campus (two men, 17 women; aged 21-60 years) to capture data for automatically labelling environment and objects (eg, fall hazards) via a CNN object detection algorithm. The case studies ran from Dec 5, 2022, to March 24, 2023. FINDINGS: To date, results show promise for the efficient, and accurate AI-based approach to better inform fall risk. Each component of the framework achieved at least 75% accuracy across a range of walks (indoor and outdoor and multiple terrains) from a dataset of 6283 new images. The AI achieved a mean average precision score of 0·93 for the identification of fall risk hazards. INTERPRETATIONS: The AI-based approach provides a contemporary means to better inform fall risk while providing an ethical means to uphold privacy. The proposed approach could have significant implications for improving overall health and quality of life, enabling ageing in place through habitual data collection with contemporary wearables to decentralise fall risk assessment. A limitation was the lack of data collection on older adults within real world, unscripted settings. However, the next phase of this research is the deployment of the AI on real-world data from a cohort of more than 40 participants within UK-based homes. FUNDING: National Institute of Health and Care Research (NIHR) Applied Research Collaboration (ARC) North-East and North Cumbria (NENC), Faculty of Engineering and Environment at Northumbria University.

Gait retraining to reduce falls: an experimental study toward scalable and personalised use in the home.

BACKGROUND: Age-related neurological conditions can result in poor mobility typified by gait abnormalities and falls, increasing risk of frailty and lowering quality of life. In the UK, the expense and inaccessibility of services to improve mobility through gait training (eg, auditory cueing) is a public health issue. Contemporary and scalable pervasive technologies for widespread public use could provide an affordable and accessible solution. We aimed to show the preliminary efficacy of a novel smartphone app that provides a personalised approach to mobility and gait assessment while facilitating gait training. METHODS: In this experimental study, we recruited participants aged 22-46 years with no physical functional impairments (ie, no age-related neurological condition and who could walk unaided) from Northumbria University staff (Newcastle upon Tyne, UK) between April 19, and May 26. Participants wore a smartphone on their lower back. Inertial data from the smartphone were recorded during two walks, one at a self-selected pace and the other with a personalised auditory cue via headphones (+10% pace on walk 1). Smartphone app functionality enabled the measurement of clinically relevant gait characteristics via a Python-based Cloud server. We compared smartphone-based mobility or gait characteristics with a gold-standard reference (Opal Mobility Lab, APDM). We used Pearson and intraclass correlation coefficients (ICC2,1) to examine agreement between the novel app and reference. The study ran from April 4 to July 21, 2023. This study received ethics approval from the Northumbria University Ethics committee, and all participants provided written informed consent. FINDINGS: Ten adults were recruited (six women and four men; mean age 27·4 years [SD 6·2], mean weight 79·6 kg [SD 12·7], mean height 174·7 cm [SD 7·9]). High levels of agreement were found between the smartphone app and reference, quantified by Pearson (≥0·858) and ICC values (≥0·911). The personalised cueing intervention increased the mean cadence by an average of 11%, which shows good participant adherence to cueing via an app. INTERPRETATION: Here, we propose a contemporary approach to increase the accessibility to a health-based intervention. Preliminary findings suggest the smartphone app is a suitable tool for personalised mobility or gait assessment while facilitating gait training. Use of a scalable app could be an accessible and affordable method for improving mobility to reduce falls in the home. Here, current limitations are the lack of investigation with the smartphone app for neurological gait assessment on older adults and the lack of information on participants app experience, but this will be included in future work. The pervasive use of smartphones enables a decentralised approach to overcoming issues such as frailty and logistical challenges of travelling to bespoke clinics. FUNDING: National Institute of Health and Care Research (NIHR) Applied Research Collaboration (ARC) North-East and North Cumbria (NENC); Faculty of Engineering and Environment at Northumbria University.

Validity and reliability of the DANU sports system for walking and running gait assessment.

Objective. Gait assessments have traditionally been analysed in laboratory settings, but this may not reflect natural gait. Wearable technology may offer an alternative due to its versatility. The purpose of the study was to establish the validity and reliability of temporal gait outcomes calculated by the DANU sports system, against a 3D motion capture reference system.Approach. Forty-one healthy adults (26 M, 15 F, age 36.4 ± 11.8 years) completed a series of overground walking and jogging trials and 60 s treadmill walking and running trials at various speeds (8-14 km hr-1), participants returned for a second testing session to repeat the same testing.Main results. For validity, 1406 steps and 613 trials during overground and across all treadmill trials were analysed respectively. Temporal outcomes generated by the DANU sports system included ground contact time, swing time and stride time all demonstrated excellent agreement compared to the laboratory reference (intraclass correlation coefficient (ICC) > 0.900), aside from ground contact time during overground jogging which had good agreement (ICC = 0.778). For reliability, 666 overground and 511 treadmill trials across all speeds were examined. Test re-test agreement was excellent for all outcomes across treadmill trials (ICC > 0.900), except for swing time during treadmill walking which had good agreement (ICC = 0.886). Overground trials demonstrated moderate to good test re-test agreement (ICC = 0.672-0.750), which may be due to inherent variability of self-selected (rather than treadmill set) pacing between sessions.Significance. Overall, this study showed that temporal gait outcomes from the DANU Sports System had good to excellent validity and moderate to excellent reliability in healthy adults compared to an established laboratory reference.

Wearables for running gait analysis: A study protocol.

Quantitative running gait analysis is an important tool that provides beneficial outcomes to injury risk/recovery or performance assessment. Wearable devices have allowed running gait to be evaluated in any environment (i.e., laboratory or real-world settings), yet there are a plethora of different grades of devices (i.e., research-grade, commercial, or novel multi-modal) available with little information to make informed decisions on selection. This paper outlines a protocol that will examine different grades of wearables for running gait analysis in healthy individuals. Specifically, this pilot study will: 1) examine analytical validity and reliability of wearables (research-grade, commercial, high-end multimodal) within a controlled laboratory setting; 2) examine analytical validation of different grades of wearables in a real-world setting, and 3) explore clinical validation and usability of wearables for running gait analysis (e.g., injury history (previously injured, never injured), performance level (novice, elite) and relationship to meaningful outcomes). The different grades of wearable include: (1) A research-grade device, the Ax6 consists of a configurable tri-axial accelerometer and tri-axial gyroscope with variable sampling capabilities; (2) attainable (low-grade) commercial with proprietary software, the DorsaVi ViMove2 consisting of two, non-configurable IMUs modules, with a fixed sampling rate and (3) novel multimodal high-end system, the DANU Sports System that is a pair of textile socks, that contain silicone based capacitive pressure sensors, and configurable IMU modules with variable sampling rates. Clinical trial registration: Trial registration: NCT05277181.

Visual Exploration While Walking With and Without Visual Cues in Parkinson's Disease: Freezer Versus Non-Freezer.

BACKGROUND: Visual cues can improve gait in Parkinson's disease (PD), including those experiencing freezing of gait (FOG). However, responses are variable and underpinning mechanisms remain unclear. Visuo-cognitive processing (measured through visual exploration) has been implicated in cue response, but this has not been comprehensively examined. OBJECTIVE: To examine visual exploration and gait with and without visual cues in PD who do and do not self-report FOG, and healthy controls (HC). METHODS: 17 HC, 21 PD without FOG, and 22 PD with FOG walked with and without visual cues, under single and dual-task conditions. Visual exploration (ie, saccade frequency, duration, peak velocity, amplitude, and fixation duration) was measured via mobile eye-tracking and gait (ie, gait speed, stride length, foot strike angle, stride time, and stride time variability) with inertial sensors. RESULTS: PD had impaired gait compared to HC, and dual-tasking made gait variables worse across groups (all P < .01). Visual cues improved stride length, foot strike angle, and stride time in all groups (P < .01). Visual cueing also increased saccade frequency, but reduced saccade peak velocity and amplitude in all groups (P < .01). Gait improvement related to changes in visual exploration with visual cues in PD but not HC, with relationships dependent on group (FOG vs non-FOG) and task (single vs dual). CONCLUSION: Visual cues improved visual exploration and gait outcomes in HC and PD, with similar responses in freezers and non-freezers. Freezer and non-freezer specific associations between cue-related changes in visual exploration and gait indicate different underlying visuo-cognitive processing within these subgroups for cue response.

Digital Fellowships: Inspiring use of contemporary technologies in applied healthcare.

The adoption of digital technologies in healthcare, accelerated by the COVID-19 pandemic, requires a well-prepared workforce capable of implementing those technologies. Here, we examine the role and impact of digital fellowships in facilitating digital transformation in healthcare systems. Digital fellowships are structured educational programmes designed to equip healthcare professionals with advanced digital skills. Focusing on UK-based initiatives like the Topol Digital Fellowship and the Fellowship in Clinical AI, we explore their efforts to prepare healthcare leaders for digital and AI adoption. Each fellowship programme provides participants with hands-on experience in digital healthcare projects and fosters interdisciplinary collaboration and post-fellowship support. We discuss how these fellowships contribute to staff retention by diversifying professional experiences and opportunities. We call for increased collaborations between universities, industry, and professional bodies to integrate lessons from digital fellowships into relevant curricula, acknowledging that digital fellowships are just one piece of the puzzle in bridging the digital skills gap in the healthcare workforce.