Dance can offer mulGple physical and mental health benefits, and community dance programs aiming to promote health and wellbeing are widely available. Such therapeuGc dance programs have been studied across populaGons, including individuals with demenGa, Parkinson's disease (PD), stroke, cancer, auGsm spectrum disorder, and neurotypical older adults (Kshtriya et al., 2015;Pa_erson et al., 2018;Takahashi et al., 2019;Bek et al., 2020;Ares-Benitez et al., 2022;Wang et al., 2022;Karkou et al., 2023). TherapeuGc dance classes can be offered online, enabling larger numbers of people to potenGally benefit, parGcularly those with limited access to in-person groups or travel-related barriers (Bek et al., 2021(Bek et al., , 2022(Bek et al., , 2025;;Ghanai et al., 2021).Health outcomes of therapeuGc dance are typically assessed using a wide variety of general and populaGonspecific measures. LimitaGons of current measurement tools may contribute to inconsistent findings in the literature, masking the true potenGal of dance to induce health benefits (Hwang and Braun, 2015;Shanahan, 2015;Bek et al., 2020;Carapelloe et al., 2020). Advances in digital technology and arGficial intelligence (AI) offer significant potenGal to obtain detailed, quanGtaGve measures of movement in both in-person and remote contexts, providing new opportuniGes to precisely evaluate therapeuGc dance outcomes. This opinion arGcle discusses how emerging technologies that can quanGfy motor and neuroplasGc changes might go beyond convenGonal outcome measures for therapeuGc dance, considers potenGal challenges of these new approaches, and outlines future direcGons for research and pracGce.Validated clinical measures are ohen used to assess physical outcomes of dance in clinical populaGons, including gait (e.g., 6-minute or 10-metre walk test), balance (Berg Balance Scale, Balance EvaluaGon Systems Test), and funcGonal mobility (Timed Up and Go, sit-to-stand) (Earhart, 2009;Ma_le et al., 2020). Measurement tools designed for specific health condiGons are also ohen used, such as the Unified Parkinson's Disease RaGng Scale (UPDRS/MDS-UPDRS) for PD (Goetz et al., 2008) and the Fugl-Meyer Assessment (Fugl-Meyer et al., 1975) for stroke. Self-report quesGonnaires may also be used to assess physical outcomes such as acGviGes of daily living, dexterity, and falls (e.g., Lee et al., 2015;Blanco-Rambo et al., 2022).These established measures can be valuable in assessing dance outcomes, and their wide use across intervenGons can facilitate comparisons between studies and meta-analyses. However, while clinical raGng scales and other convenGonal assessment tools have yielded much useful data on dance outcomes, they are not designed to capture the specific effects of dance. Some further limitaGons of these tools are considered below.SubjecGve raGng scales are limited in reliability and sensiGvity (e.g., Chan et al., 2018;Hsiao et al., 2021), and some (e.g., UPDRS) require trained professionals to administer. PaGent-reported outcomes are valuable to understand parGcipants' experiences from their own perspecGve, but may be impacted by cogniGve impairment or reduced insight, parGcularly in neurological condiGons (e.g., Cameron et al., 2024).ExisGng technology-based measures of dance outcomes include instrumented gait mats, treadmills, balance boards, and force plates (e.g., Guzmán-García et al., 2011;Sohn et al., 2018) as well as automated versions of established clinical tests, such as Timed Up and Go and sit-to-stand (Tan et al., 2019;Vourganas et al., 2019). While gait, funcGonal mobility, and balance are important outcomes in terms of falls prevenGon and mobility, it is also useful to assess upper-body kinemaGcs, including limb movements and fine motor coordinaGon, which may be improved through dance (parGcularly in seated dance programs designed for people with limited mobility). PotenGal upper-body outcomes of dance that are not well addressed by current measures include bimanual coordinaGon, sequencing, amplitude, smoothness, and range of moGon.In summary, the beneficial effects of dance may be underesGmated because many current measurement techniques are not sufficiently sensiGve to change or do not fully capture relevant aspects of movement addressed by dance.These limitaGons are further compounded by small sample sizes (e.g., Pa_erson et al., 2018;Bek et al., 2020b;Ares-Benitez et al., 2022). AddiGonally, few studies have tracked outcomes of therapeuGc dance over an extended period to assess ongoing improvement or maintenance of benefits (for reviews, see Sharp and Hewi_, 2014;Kalyani et al., 2019;Bek et al., 2020b).Technological developments, parGcularly in remote moGon capture and mobile neuroimaging, expand the possibiliGes for precise and reliable measurement of therapeuGc benefits, potenGally overcoming some of the limitaGons discussed above. As technology advances, the ability to measure outcomes not only in controlled clinical or laboratory environments but also in real-world daily living contexts is rapidly expanding.MoGon capture and wearable moGon sensors (inerGal measurement units; IMUs) have long been used to assess outcomes of intervenGons, although few studies used these to test the effects of dance (Hulbert et al., 2020;Haputhanthirige et al., 2023). These technologies enable more comprehensive measurement of full-body movement than convenGonal laboratory-based instruments such as gait mats and balance boards.TradiGonal laboratory-based moGon capture systems, which uGlize reflecGve or infrared markers, are expensive and require Gme and experGse for setup and analysis. Portable moGon capture using depth sensing systems like Kinect (e.g., Sampaio et al., 2016;Garcia-Agundez et al., 2019;Lu et al., 2024), as well as virtual reality (Lee et al., 2015;Bok et al., 2023) or newer approaches such as LiDAR (Yoon et al., 2021) can be used in the home or community to deliver therapeuGc programs and monitor outcomes. However, these methods may have lower spaGal and temporal resoluGon or require addiGonal processing compared to emerging techniques. Newer video-based markerless moGon capture approaches can use computer vision and machine learning to esGmate human joint posiGons directly from video, providing powerful tools that non-experts can set up and use. Markerless moGon capture using computer vision can provide detailed kinemaGc measures of gait, posture, and coordinaGon from a simple camera setup. This type of technology has already been used to study movement in people with PD, stroke and cerebral palsy in non-dance contexts (MarGnez et al., 2018;Steffensen et al., 2023). Moreover, while laboratory-based markerless systems -such as Theia3D (Theia Markerless Inc., Kingston, ON, Canada) -ohen sGll rely on expensive licensed sohware, open-source sohware for moGon capture and analysis can remove physical and financial barriers to technology-derived measurement. For example, recent studies in PD have used computer vision analysis (e.g., DeepLabCut, Mediapipe) of pre-recorded videos of hand movements to classify disease state (Heye et al., 2024) and to idenGfy kinemaGc markers indicaGng response to levodopa treatment (Lange et al., 2025). These newer moGon capture technologies also enable greater flexibility to measure outcomes remotely and at scale.ExisGng methods using wearable inerGal sensors, smartphones, fitness trackers, and smartwatches can capture movement intensity, gait, and balance during dance acGviGes and daily living (Blackler et al., 2019;Avci, 2024; for review see Tao et al., 2024). In addiGon, wearable sensors can capture data related to physical acGvity, falls, sleep pa_erns, and other physiological indicators of health (e.g., Kristoffersson and Lindén, 2022;Ghazi et al., 2025). However, while most wearable devices currently available are rigid, visible, and can be uncomfortable to wear -limiGng their acceptability -the advent of soh smart wearable sensors that conform to the skin and stretch with the body will likely make these technologies more pracGcal and appealing (Kim et al., 2025), thereby increasing the feasibility of longitudinal data collecGon to invesGgate longer-term effects of dance.Neuroimaging methods can be used to detect neuroplasGc changes resulGng from therapeuGc dance in motorrelated and other brain regions, complemenGng behavioural outcome measures. MagneGc resonance imaging (MRI) techniques such as fMRI and diffusion tensor imaging (DTI) have been used to show funcGonal or structural connecGvity changes in the brain associated with dance (Teixeira-Machado et al., 2019;Meulenberg et al., 2023;Simon et al., 2024;Tung et al., 2024;Wu et al., 2025). For example, a case study of an individual with PD parGcipaGng in regular dance classes found funcGonal changes in brain areas involved in movement planning and imagery, rhythm, emoGonal processing, and mulGsensory integraGon (Simon et al., 2024). In addiGon, dance-based intervenGons in older adults have been associated with increases in integrity of white-ma_er tracts and changes in structural connecGvity in networks relevant to motor control, balance, and coordinaGon. For example, a study of aerobic dance in older adults with mild cogniGve impairment showed enhanced structural connecGons within the default mode network and between the supplementary motor area and default mode network regions (Wu et al., 2025).Emerging neuroimaging technologies now have the capacity to record neural acGvity not only before and aher, but also during dance. One such method, funcGonal near-infrared spectroscopy (fNIRS), allows measurement of corGcal hemodynamics during movement, enabling more direct invesGgaGon of therapeuGc mechanisms. For example, an 8-week interacGve dance training study in older adults found that changes in prefrontal cortex oxygenaGon during treadmill walking correlated with improvements in execuGve funcGoning (Eggenberger et al., 2016). Mobile electroencephalography (EEG) approaches similarly enable real-Gme assessment of sensorimotor neural rhythms and event-related desynchronizaGons during walking (Bonassi et al., 2024), which can also provide insight into mechanisms underlying therapeuGc effects of dance, parGcularly when combined with kinemaGc measures or correlated with physical funcGon outcomes. Mobile brain/body imaging (MoBI; Barnstaple et al., 2021;King and Parada, 2021) is an example of such an approach, integraGng EEG with moGon capture during acGve movement.The integraGon of new technologies into dance outcome measurement presents several challenges. The rapid pace of digital innovaGon and AI advances will require ongoing evaluaGon and updaGng of assessment tools to opGmize validity and reliability. A key issue is the need to establish technology-derived outcomes that are both clinically meaningful and populaGon-specific and can funcGon as biomarkers of therapeuGc effects (i.e., kinemaGc or neural signatures of dance outcomes).AddiGonally, clinicians and researchers may be hesitant to move away from gold standards, while parGcipants may have reservaGons about being monitored or having their data collected remotely. Remote assessments also depend on digital literacy, internet connecGvity, and access to devices (e.g., Bek et al., 2021;Okafor et al., 2024), and may sGll require healthcare professionals to assist with setup and training. Finally, cultural differences, such as in aetudes toward technology and privacy, must be considered.Future research direcGons will be partly driven by ongoing technological advances. Nonetheless, we can consider some ways in which emerging technologies may contribute to enhancing quanGtaGve outcome measurement in therapeuGc dance. With further research and validaGon, technology-derived measures could idenGfy kinemaGc or neural "signatures" of improvement in different populaGons. Remote data collecGon will enable larger-scale studies to be conducted using home-based training. Open-source sohware and code sharing could facilitate greater standardizaGon of measurement, leading to more meaningful comparison of outcomes. Moreover, combining technologies such as moGon capture and mobile imaging will enable researchers to idenGfy neural correlates of motor funcGon in real-Gme, to be_er understand brain mechanisms underlying physical outcomes. Beyond gathering an evidence base to support the development of dance programs and intervenGons, digital technologies and AI also offer the opportunity to provide personalized feedback and tailored progression informed by dynamic analysis of performance data (e.g., Kang et al., 2023;Liu et al., 2023).The technologies discussed here in relaGon to dance can also be applied across many other therapeuGc contexts. FacilitaGng self-monitoring and independent home-based training, where appropriate, can empower paGents to take a more acGve role in managing their rehabilitaGon (e.g., Kraal et al., 2013;Doyle et al., 2019). AddiGonally, data can be curated and shared with clinicians and healthcare providers to help guide care.Further work should cross-validate new measures with gold standards to provide evidence of convergent validity and facilitate meaningful interpretaGon of novel metrics. For example, computer vision measures of bradykinesia have been validated against clinical raGngs of PD symptoms from the MDS-UPDRS (Heye et al., 2024). While digital technology-enabled methods may not replace clinical evaluaGons and paGent-reported outcomes, which will remain important tools in interpreGng quanGtaGve metrics relaGve to contextual factors, they will expand the capacity of clinicians and researchers to precisely, comprehensively, and efficiently evaluate therapeuGc dance programs and deliver personalized training.
Source
Judith Bek; Deborah A. Jehu; Gammon M. Earhart; Madeleine E. Hackney. Frontiers in Psychology, 2026. DOI: 10.3389/fpsyg.2026.1749839