Review Personalized Interactive Music Systems for Physical Activity and Exercise: Exploratory Systematic Review and Meta- Analysis Andrew Danso1, PhD; Tiia Kekäläinen2,3, PhD; Friederike Koehler1, PhD; Keegan Knittle4, PhD; Patti Nijhuis1, PhD; Iballa Burunat1, PhD; Pedro Neto1, PhD; Anastasios Mavrolampados1, PhD; William M Randall1, PhD; Niels Chr Hansen1,5,6, PhD; Alessandro Ansani1, PhD; Timo Rantalainen4, PhD; Vinoo Alluri7, PhD; Martin Hartmann1, PhD; Rebecca S Schaefer8,9,10, PhD; Johanna K Ihalainen4,11, PhD; Rebekah Rousi12, PhD; Kat R Agres13, PhD; Jennifer MacRitchie14, PhD; Petri Toiviainen1, PhD; Suvi Saarikallio1, PhD; Sebastien Chastin15,16, PhD; Geoff Luck1, PhD 1Department of Music, Arts and Culture Studies, Centre of Excellence in Music, Mind, Body and Brain, University of Jyväskylä, Jyväskylä, Finland 2Laurea University of Applied Sciences, Vantaa, Finland 3Gerontology Research Center and Faculty of Sport and Health Sciences, University of Jyväskylä, Jyväskylä, Finland 4Faculty of Sport and Health Sciences, University of Jyväskylä, Jyväskylä, Finland 5Royal Academy of Music Aarhus, Aarhus, Denmark 6Cognitive Musicology and Performance Science Lab, Department of Communication and Psychology, Aalborg University, Aalborg, Denmark 7Cognitive Science Lab, International Institute of Information Technology, Hyderabad, India 8Health, Medical, and Neuropsychology Unit, Institute of Psychology, Faculty of Social and Behavioural Sciences, Leiden University, Leiden, The Netherlands 9Leiden Institute for Brain and Cognition, Leiden University, Leiden, The Netherlands 10Academy of Creative and Performing Arts, Faculty of Humanities, Leiden University, Leiden, The Netherlands 11Finnish Institute of High Performance Sport KIHU, Jyväskylä, Finland 12School of Marketing and Communication, Communication Studies, University of Vaasa, Vaasa, Finland 13Centre for Music and Health, Yong Siew Toh Conservatory of Music, National University of Singapore, Singapore, Singapore 14Department of Music & Healthy Lifespan Institute, University of Sheffield, Sheffield, United Kingdom 15School of Health and Life Sciences, Glasgow Caledonian University, Glasgow, United Kingdom 16Department of Movement and Sports Sciences, Ghent University, Ghent, Belgium Corresponding Author: Andrew Danso, PhD Department of Music, Arts and Culture Studies Centre of Excellence in Music, Mind, Body and Brain, University of Jyväskylä Seminaarinkatu 15, Jyväskylän yliopisto Jyväskylä 40014 Finland Phone: 358 6643034 Email: andrew.a.dansoadu@jyu.fi Abstract Background: Personalized Interactive Music Systems (PIMSs) are emerging as promising devices for enhancing physical activity and exercise outcomes. By leveraging real-time data and adaptive technologies, PIMSs align musical features, such as tempo and genre, with users’ physical activity patterns, including frequency and intensity, enhancing their overall experience. Objective: This exploratory systematic review and meta-analysis evaluates the effectiveness of PIMSs across physical, psychophysical, and affective domains. Methods: Searches across 9 databases identified 18 eligible studies, of which 6 (comprising 17 intervention arms) contained sufficient data for meta-analysis. Random-effects meta-analyses and meta-regression were performed to assess outcomes for physical activity levels, physical exertion, ratings of perceived exertion, and affective valence. Results: Results showed significant improvements in physical activity levels (g=0.49, CI 0.07 to 0.91, P=.02, k=4) and affective valence (g=1.65, CI 0.35 to 2.96, P=.01, k=4), with faster music tempo identified as a significant moderator (P=.03). JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 1 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 No significant effects were observed for ratings of perceived exertion (g=0.72, CI −0.13 to 1.58, P=.10, k=3) or physical exertion (g=0.78, CI −0.55 to 2.11, P=.25, k=5). Conclusions: Substantial heterogeneity and limited study quality indicate the need for more robust, randomized controlled trials to establish the efficacy of PIMSs in diverse populations. Trial Registration: PROSPERO (International Prospective Register of Systematic Review) CRD42023465941; https:// www.crd.york.ac.uk/PROSPERO/view/CRD42023465941 JMIR Hum Factors 2025;12:e70372; doi: 10.2196/70372 Keywords: music intervention; health promotion; exercise; affect; systematic review; meta-analysis; mobile phone Introduction Background Regular physical activity and exercise are fundamental to maintaining and enhancing overall health and well-being. Despite their recognized role in preventing and managing noncommunicable diseases such as cardiovascular diseases, cancer, and diabetes, engagement in regular physical activity and exercise remains below the suboptimal level [1]. This deficiency undermines the potential for mental health benefits of physical exercise and its contributions to quality of life [2]. The World Health Organization defines physical activity broadly, encompassing all forms of bodily move- ment generated by skeletal muscles that require energy expenditure, including activities such as walking, sports, and dance [1]. In contrast, exercise has been defined as “physical activity that is planned, structured, repetitive, and purposive, aiming to improve or maintain one or more components of physical fitness” [3]. However, the broad spectrum of activities categorized as physical activity and exercise often presents challenges in promoting consistent engagement and uptake, including individual-level barriers such as motivation and time constraints [4]. Efforts to increase engagement in physical activity and exercise have faced significant challenges, frequently yielding inconsistent outcomes, as exemplified by interventions such as pedome- ter-based programs, which demonstrate variable effectiveness depending on factors including participant motivation and engagement [4,5]. Role of Music in Enhancing Physical Activity and Exercise Music’s rhythmic properties have been shown to influ- ence perceptions, ergonomics, and physiological markers associated with physical activity and exercise [6-10]. Available evidence suggests that auditory-motor coupling facilitates predictive synchronization in physical activity and exercise settings, which can reduce perceived exertion and enhance endurance [11,12]. Additionally, when music aligns with individual preferences, such as through self-selection, it may further increase motivation, improve affective states, induce distraction, and lower perceived effort during physical activities and exercise [7,11]. The integration of music into exercise contexts can be further understood through theoretical frameworks such as the Affective-Reflective Theory (ART) and Dual-Mode Theory. ART emphasizes the importance of momentary affective responses—such as pleasure or displeasure—in shaping future exercise behaviors [13,14]. These responses, encap- sulated in the construct of “affective valence,” reflect the intrinsic pleasantness or unpleasantness of emotional states that fluctuate in response to internal and external stimuli. Conversely, Dual-Mode Theory posits that music’s impact on affective responses is most pronounced at moderate exercise intensities, within a zone of response variability. This zone refers to the range of exercise intensity where affective responses—such as feelings of pleasure or displeas- ure—are particularly sensitive to individual differences (eg, fitness level and psychological state) and contextual factors (eg, music, environment, and social setting). In this range, attentional focus and physiological cues mediate affective experiences [15]. While both theories acknowledge the importance of affective responses in exercise, Dual-Mode Theory provides a more nuanced perspective by emphasiz- ing intensity-dependent variability and its interaction with individual and contextual factors. Extending the principles of ART and Dual-Mode Theory [16], the framework highlights how music’s intrinsic properties—such as tempo, rhythm, and harmony—interact with personal and situational moderators, including exercise intensity and individual preferences, to influence affective and behavioral outcomes in exercise. Music operates through 3 primary mechanisms: regulating affective states, disso- ciating attention from exertional discomfort, and facili- tating temporal prediction and rhythmic synchronization. These mechanisms are most effective within the zone of response variability, where affective valence dynamically influences exercise engagement [15,17]. Empirical studies consistently demonstrate that personalized music enhances energy efficiency, reduces perceived exertion, and improves adherence by fostering positive affect [17]. Such evidence positions personalized music systems as a key tool for optimizing both the immediate and long-term benefits of exercise. Personalized Interactive Music Systems in Physical Activity and Exercise Recently, advances in personalized music technologies have led to the development of Personalized Interactive Music Systems (PIMSs), which leverage software, sensors, and computer algorithms to deliver a dynamic, tailored music experience during physical activity and exercise [18,19]. These systems integrate with smartphones and wearable JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 2 (page number not for citation purposes) https://doi.org/10.2196/70372 https://humanfactors.jmir.org/2025/1/e70372 devices to monitor user movements and adjust musical features, such as tempo, style, and timbre, in real time to align with exercise routines, enhancing engagement and adherence to activity [20,21]. PIMSs have been designed for diverse contexts, targeting both intrinsic factors (eg, motivation and attentional focus) and extrinsic factors (eg, training guidance). For example, a PIMS, the moBeat system, used real-time interactive music and biophysical feedback to enhance cycling performance by increasing intrinsic motivation and maintaining pace and intensity [12]. Similarly, PIMS interventions for older adults have demonstrated benefits for physical endurance and engagement relative to conventional exercise conditions [22]. As mobile interventions incorporating personalization have been shown to be more effective at enhancing physical activity than nonpersonalized approaches [23], PIMSs hold promise for improving physical activity adherence, reducing the ratings of perceived exertion (RPE), and fostering positive affective states during exercise by dynamically tailoring music to individual physiological, affective, and contextual needs [7,12,22]. Due to the relatively recent advancements of PIMSs, there is yet limited empirical evidence on their effective- ness across physical activity and exercise-related domains. Such information is essential for informing implementa- tion, replication, and comparative evaluation of interven- tions aimed at promoting adherence to physical activity and exercise [24]. While systematic reviews and meta-anal- yses have explored the general effects of music on phys- ical activity and exercise-related outcomes [6,7,25], these reviews predominantly focused on traditional music listening interventions and did not systematically evaluate the impact of personalized and interactive music systems. By specifically examining PIMSs, this review and meta-analysis contribute to understanding how tailored, interactive music interventions influence physical, psychophysical, and affective dimen- sions of physical activity and exercise engagement, thereby addressing a critical gap in the existing literature. Therefore, this study combines a systematic review and exploratory meta-analysis to evaluate the effectiveness of PIMSs on physical activity and exercise-related outcomes. Specifically, this study synthesizes findings on physical activity levels, psychophysical measures (eg, RPE and physical exertion), and affective outcomes (eg, affective valence and mood states). Our main research question is: How effective are PIMSs across physical, psychophysical, and affective outcomes during physical activity and exercise? This analysis intends to provide early insights into the specificity of PIMSs’ effects and identify gaps in the literature that warrant further investigation. Methods Study Design This systematic review and meta-analysis were designed based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) protocol [26]. The full search strategy can be found in the review registration document (CRD42023465941). Eligibility Criteria We included (1) studies investigating the effect of PIMSs on physical activity or exercise, including their effects on motivation, exercise intensity, adherence, or related outcomes, (2) studies including participants from diverse populations (eg, sufficiently active and not sufficiently active individuals), and (3) papers in the English language, published from January 2010 to May 2024 in peer-reviewed journals or as published proceedings (conference papers were considered due to the limited number of peer-reviewed studies). We excluded (1) studies from nonpeer-reviewed sour- ces, books, dissertations, and theses; (2) papers written in languages other than English; and (3) studies that were not directly related to the effect of PIMSs on physical exercise or physical activity. Information Sources We searched the following databases: (1) Web of Science, (2) SPORTDiscus, (3) Medline, (4) Embase, (5) ACM Digital Library databases, (6) Springer, (7) Google Scholar, (8) IEEE Xplore, and (9) Scopus. The database search was supplemen- ted by a backward snowball search, whereby the reference list of all papers was scanned for potential sources. The snow- ball search continued until no new sources could be identi- fied. The initial interrater agreement for the identification of relevant sources was k=0.83, indicating a strong level of agreement among the 2 individuals performing 2 independent snowball searches (AD and TK). Full search strings for all databases used in this review are provided in Section S3 in Multimedia Appendix 1. Search Strategy A literature search was performed using terminology related to the effects of PIMSs on physical activity and exer- cise, (“Personali*ed Interactive Music System*” OR “Music Recommendation Algorithm” OR “Music Recommendation System*” OR “Streaming” OR “MP3” OR “Digital Music”) AND (“Physical Activity” OR “Exercise” OR “Recovery” OR “Recuperation” OR “Sedentary Behav*” OR “Physical Inactivity”). Selection Process and Data Collection Process The citations of all retrieved papers were imported into Zotero (Digital Scholar), where duplicates were systemati- cally identified and removed. Subsequently, 2 authors (AD and TK) independently screened the titles and abstracts of JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 3 (page number not for citation purposes) https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=465941 https://humanfactors.jmir.org/2025/1/e70372 the studies using ASReview (Utrecht University) [27] and Rayyan [28]. Papers that could not be definitively excluded based on the title or abstract underwent full-text retrieval for further evaluation. The full-text papers were then independ- ently assessed for inclusion by the same 2 authors (AD and TK). Disagreements at any stage were resolved through discussion, with a third author consulted to achieve consensus when necessary. Data Extraction The studies’ information was extracted to a spreadsheet, including study characteristics, such as the type of PIMS, the study design, PIMS measurement, and the target behavior of the PIMS (Target Physical Activity or Exercise; Table 1). JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 4 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 Ta ble 1. C ha rac ter ist ics of in clu de d s tud ies . Re fer e nc e Co un try Ag e ( ye ars ) Sa mp le siz e (N ) Po pu lat ion Ty pe of PI M Sa Stu dy de sig n PI M S me asu rem en t Ta rge t b eh av ior or tar ge t p hy sic al ac tiv ity or ex erc ise Ph ys ica l ac tiv ity res ult s G (95 % CI ) o f P IM S o n o utc om es of int ere st [29 ] Ca na da • 47 .3‐ 79 .2b • 34 Pa tie nts w ith ca rdi ov asc ula r dis ea se Pe rso na liz ed mu sic au dio - pla yli sts Ra nd om ize d ex pe rim en tal de sig n Tr iax ial ac ce ler om e- ter Ad he ren ce Im pro ve d P Ac vo lum es (P <.0 01 ) • g= 0.5 1 ( −0 .47 to 1. 49 ) f or ph ys ica l a cti vit y lev el (R AS d ) • g= −0 .06 (− 1.0 4, 0.9 2) for ph ys ica l a cti vit y lev el (no R AS ) [30 ] Sp ain • N/ Ae • N/ A N/ A Pe rso na liz ed mu sic rec om me nd ati on sy ste m Pr oo f o f c on ce pt Se ns ors f M oti va tio n o r pe rfo rm an ce en ha nc em en t N/ A • N/ A [31 ]g Ge rm an y • N/ A • 1 Ol de r a du lt pa rti cip an t M us ic fee db ac k for re ha bil ita tio n Pr oo f o f c on ce pt Ac ce ler om ete r Re ha bil ita tio n N/ A • N/ A [32 ]g Ta iw an • 21 .56 (S D 1.0 4)h • 10 fe ma le an d 2 6 ma le Pa rti cip an ts fro m the Na tio na l Y an g M ing C hia o Tu ng Un ive rsi ty Ex erc ise sy ste m for m idd le- dis tan ce ru nn ing Ex pe rim en tal Sm art ph on e’s bu ilt- in tri ax ial ac ce ler om ete r Ad ap tin g m us ic sel ec tio n t o t he us er’ s pa ce du rin g w alk ing N/ A • g= −0 .73 (− 1.4 0 t o − 0.0 6) for ph ys ica l e xe rti on • g= 1.6 3 ( 0.8 8 t o 2 .39 ) f or RP Ei • g= 2.1 7 ( 1.3 4 t o 2 .99 ) f or aff ec tiv e v ale nc e [33 ] Ta iw an • N/ A • N/ A N/ A M us ic ass ist ed run tr ain er Pr oo f o f c on ce pt Tr iax ial ac ce ler om ete r Ph ys iol og ica l, pe rce ptu al, an d aff ec tiv e r esp on ses N/ A • N/ A [34 ]g Sin ga po re • N/ A • 60 Stu de nts A mu sic rec om me nd ati on sy ste m W ith in- su bje cts cro sso ve r d esi gn M us ic rec om me nd ati on rat ing s M oti va tio n N/ A • N/ A [35 ] N/ A • N/ A • 45 No na thl ete s, no nb od y bu ild ers , no nm us ici an s JY M M iN se ns or att ac he d t o fit ne ss de vic es to pro vid e mu sic al fee db ac k Ex pe rim en tal de sig n M ov em en t s en so rj W ork ou t N/ A • g= 0.9 6 ( 0.3 5 t o 1 .56 ) f or aff ec tiv e v ale nc e [36 ]g N/ A • N/ A • 27 N/ A Ru nn er’ s Ju ke bo x: mu sic tem po m atc hin g the us er’ s p ac e du rin g e xe rci se Us er tes tin g d esi gn Sm art ph on e a pp to rec og niz e u ser pa ce or ad jus t m us ic tem po W alk ing or ru nn ing pa ce m on ito rk N/ A • g= 2.7 6 ( 1.6 3 t o 3 .89 ) f or ph ys ica l e xe rti on (fi xe d B PM l ) • g= 2.2 9 ( 1.2 4 t o 3 .34 ) f or ph ys ica l e xe rti on (pa ce -m atc he d) • g= −0 .74 (− 1.6 1 t o 0 .13 ) f or ph ys ica l e xe rti on (ra nd om ) [37 ]g De nm ark • N/ A • N/ A N/ A M us ic cli ps w ith dy na mi c B PM ran gin g f rom 11 0‐ 17 0 Pr oo f o f c on ce pt Se ns ors m Cy cli ng N/ A • N/ A [38 ] Sw itz erl an d • 18 ‐4 5 • 7 f em ale s an d 8 ma les Cy cli sts So un dB ike : mu sic al so nif ica tio n t o im pro ve sp on tan eo us sy nc hro niz ati on of cy cli sts Ex pe rim en tal Se ns ors Cy cli ng En ha nc ed cy cli st sy nc hro niz ati on to ex ter na l mu sic • N/ A   JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 5 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372   Re fer e nc e Co un try Ag e ( ye ars ) Sa mp le siz e (N ) Po pu lat ion Ty pe of PI M Sa Stu dy de sig n PI M S me asu rem en t Ta rge t b eh av ior or tar ge t p hy sic al ac tiv ity or ex erc ise Ph ys ica l ac tiv ity res ult s G (95 % CI ) o f P IM S o n o utc om es of int ere st [39 ] Fin lan d • N/ A • 2 Ol de r a du lt pa rti cip an ts Pr oc ess ing ac ce ler om etr y da ta to cre ate mu sic al so nif ica tio ns of ph ys ica l a cti vit y Pr oo f o f c on ce pt So nif ica tio n o f P A da tan Aw are ne ss of PA N/ A • N/ A [40 ]g Be lgi um • N/ A • 33 Pa rti cip an ts fro m pu bli c ev en t DS aT o alg ori thm fo r mu sic se lec tio n an d r ea l-t im e ad ap tat ion Pil ot stu dy Tr iax ial ac ce ler om ete r Sy nc hro niz ati on to th e be at of mu sic Th e m ajo rit y (19 /33 , 5 8% ) sy nc hro niz ed the ir ste ps w ith mu sic • N/ A [41 ] Be lgi um , Cz ec h Re pu bli c • 21 .9 (S D 12 .9) h • 20 .2 (S D 0.8 )h • 21 .2 (S D 1.7 )h • 23 (S D 3)h • 82 m ale an d 6 8 fem ale • 56 m ale an d 4 4 fem ale • 12 fe ma le • 6 f em ale an d 4 ma le N/ A Sy nc hro niz e mu sic w ith th e pa rti cip an t’s mo ve me nts Ca se stu dy Re co rdi ng s o f foo tfa lls an d m us ic ali gn me nt str ate gie sp Sy nc hro niz ati on to th e be at of mu sic Im pro ve d en tra inm en t • N/ A [42 ]g N/ A • N/ A • N/ A N/ A Co nte xt- aw are rec om me nd er sy ste m M ixe d m eth od s de sig n q Au tom ati c l ea rni ng alg ori thm M oti va te us ers to co mp let e P A N/ A • N/ A [22 ] Ge rm an y • 70 .6 (S D  3.9 ) h • 11 fem ale s an d 5 ma les No np hy sic all y ac tiv e JY M M iN : sen so r a tta ch ed to fit ne ss de vic es to pro vid e m us ica l fee db ac k W ith in- su bje cts de sig n M ov em en t s en so rj Str en gth -en du ran ce ex erc ise s N/ A • g= 0.7 3 ( 0.0 0 t o 1 .46 ) f or ph ys ica l a cti vit y l ev el • g= 0.2 0 ( −0 .27 to 0. 67 ) f or RP E • g= 0.0 9 ( −0 .47 to 0. 64 ) f or aff ec tiv e v ale nc e [43 ]g Ne the rla nd s • 18 ‐2 5 • 24 Of fic e wo rke rs Sm art cu sh ion pro vid ing mu sic al fee db ac k W ith in- su bje cts de sig n M ov em en t s en so r p ad Po stu re ch an ge s No ef fec t o n bre ak ing sed en tar y be ha vio r • N/ A [44 ] No rw ay • N/ A • 3-6 r Se nio rs wi th ea rly -st ag e Al zh eim er dis ea se Int era cti ve mu sic sy ste m Qu ali tat ive res ea rch de sig n Se ns or pa d Sti mu lat e o r m oti va te PA N/ A • N/ A [12 ] Ne the rla nd s • 23 ‐5 1 • 26 Ph ilip s em plo ye es mo Be at: int era cti ve mu sic sy ste m W ith in- su bje ct ex pe rim en t Ca de nc e s en so r, h ea rt rat e M oti va tio n N/ A • g= 0.5 4 ( −0 .22 to 1. 30 ) f or ph ys ica l a cti vit y lev el • g= 0.5 4 ( −0 .22 to 1. 30 ) f or ph ys ica l e xe rti on • g= 0.4 3 ( −0 .32 to 1. 19 ) f or RP E   JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 6 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372   Re fer e nc e Co un try Ag e ( ye ars ) Sa mp le siz e (N ) Po pu lat ion Ty pe of PI M Sa Stu dy de sig n PI M S me asu rem en t Ta rge t b eh av ior or tar ge t p hy sic al ac tiv ity or ex erc ise Ph ys ica l ac tiv ity res ult s G (95 % CI ) o f P IM S o n o utc om es of int ere st • g= 3.7 9 ( 2.5 2 t o 5 .06 ) f or aff ec tiv e v ale nc e a P IM S: Pe rso na liz ed In ter ac tiv e M us ic Sy ste m. b L ow est lo we r b ou nd : 4 7.3 ye ars (f rom th e s ec on d s ub gro up ). H igh est up pe r b ou nd : 7 9.2 ye ars (f rom th e f irs t s ub gro up ). T he es tim ate d e nti re ag e r an ge fo r a ll 3 gr ou ps co mb ine d w ou ld be fr om ap pro xim ate ly 47 .3 to 79 .2 ye ars . c P A: ph ys ica l a cti vit y. d R AS : r hy thm ic au dit ory st im ula tio n. e N /A : n ot ap pli ca ble . f G alv an ic sk in res po ns e, ox yg en sa tur ati on se ns or, an d p uls e s en so r. g C on fer en ce pa pe rs. h M ea n ( SD ). i R PE : r ati ng s o f p erc eiv ed ex ert ion . j JY M M iN : th e m ov em en t o f t he se ns or- eq uip pe d f itn ess de vic e i s m ap pe d t o m us ica l p ara me ter s, cre ati ng an ac ou sti c f ee db ac k s ign al. k S W PM : s wi ng s p er mi nu te. l B PM : b ea ts pe r m inu te. m M on ito r c yc lin g p ac e a nd he art ra te, in flu en cin g a ud io fee db ac k ( so un ds ca pe so un ds ) i n r ea l-t im e. n A cc ele rom etr y d ata . o D Sa T: D yn am ic So ng an d T em po . p T he m eth od olo gy in vo lve d r ec ord ing fo otf all s a nd va rio us m us ic ali gn me nt str ate gie s t o s yn ch ron ize m us ic wi th pa rti cip an ts’ w alk ing or ru nn ing m ov em en ts. q In clu de s e lem en ts of a p roo f o f c on ce pt de sig n a nd an ex pe rim en tal de sig n. r E xa ct nu mb ers ar e n ot sp ec ifi ed , b ut a m en tio n o f a gr ou p s ize of 3 to 6 p art ici pa nts . JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 7 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 Preregistration Deviations Where available, quantitative data suitable for meta-analysis were extracted. This was done for the preregistered outcome of physical activity level, as well as for affective valence, RPE, and physical exertion, which were not preregistered as outcomes. The decision to extract data on these addi- tional outcomes was taken because of the close relationships between these variables and physical activity and exercise participation, their prevalence as outcomes in the included studies, and the limited number of studies reporting data on physical activity and exercise behavior. In cases where effect sizes could not be readily calculated based on the published papers, their authors (n=2) were contacted at least twice for additional data, resulting in the provision of calculations for 5 additional effect sizes. Operationalization of Terms This review operationalizes 4 key terms central to physi- cal activity and exercise research. Physical activity level is defined by the quantified volume (eg, daily activity counts and weekly minutes), intensity (eg, metabolic equivalent of task [MET] and percent oxygen uptake reserve), and compliance (eg, adherence to heart rate zones or regimens) [45,46] of physical activity. Affective valence refers to the pleasure-displeasure dimension of emotional responses during or after physical activity, assessed using self-report scales such as the Feeling Scale (FS) [47] and the “good versus bad mood” subscale of the Multidimensional Mood Questionnaire (MDMQ) [48]. These measures capture subjective ratings of positivity or negativity without incorporating arousal [14,35,49]. Physical exertion encompasses physiological (eg, heart rate), biomechanical (eg, stride length), and perceptual demands, providing a comprehensive assessment of effort [50]. These constructs serve as the primary outcomes of interest in this review. The constructs are summarized in Table 2 and described further in Section S1 in Multimedia Appendix 1. Table 2. Operationalization of terms. Term Definition Operational metrics References Physical activity level Encompasses the volume, intensity, and compliance with physical activity recommen- dations or exercise regimens. • Volume: total activity counts per day via accelerometer, mean weekly minutes. Intensity: absolute intensity using metabolic equivalent of tasks, relative intensity as oxygen consumption reserve percentage. • Compliance: adherence to recommendations or regimens via changes in volume, device usage, or adherence to heart rate zones. [45,46] Physical exertion Effort exerted to perform physical activity, involving physiological, biomechanical, and perceptual demands. • Physiological: heart rate as an indicator of cardiovascular response. • Biomechanical: stride length and pace for activities such as running and walking. • Perceptual: integration of physiological and biomechanical cues to assess overall effort. [50] RPEa Subjective numerical value reflecting perceived effort during physical activity, integrating sensory cues, and physiological sensations. • Scale: Borg RPE scale for aerobic activities (cycling and running). • Category-Ratio Scale: Borg Category-Ratio 10 Scale to measure perceived exertion or other sensations. • Responses: local sensations (muscles, skin, and joints) and central factors (cardiopulmonary system). [7,51,52] Affective valence The subjective feeling of pleasure or displeasure experienced during or after physical activity. It is independent of perceived exertion and reflects emotional responses to exercise, influenced by individual, contextual, and social factors. Affective valence is measured using self-report scales, such as: • Scale: ○ Feeling Scale: bipolar scale from +5 (very good) to −5 (very bad). ○ Positive and negative affect schedule: assesses positive and negative emotions. ○ Multidimensional Mood Questionnaire (MDMQ): evaluates mood during exercise using subscales for “good versus bad mood,” “calmness versus agitation,” and “alertness versus tiredness.” Only the “good versus bad mood” subscale aligns with the pleasure-displeasure dimension of affect. • Context: Measurement occurs before, during or immediately after exercise. [14,47-49] aRPE: ratings of perceived exertion. JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 8 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 Study Risk of Bias Assessment The quality of the studies was assessed by 2 authors (AD and TK) using the JBI’s (Joanna Briggs Institute) critical appraisal checklist, including tools for quasi-experimental appraisal, qualitative research appraisal, and the revised checklist for randomized controlled trials [53]. Data Synthesis and Analysis Methods We conducted a narrative synthesis, categorizing studies into two groups based on design: (1) experimental stud- ies, including randomized, quasi-experimental, pilot, and within-subject designs, and (2) proof-of-concept and user- testing studies. This classification enabled the identification of trends within and across these categories. For experimental studies, we examined outcomes related to physical activity levels, physical exertion, RPE, and affective valence. Proof-of-concept and user-testing studies were analyzed for their focus on PIMS design features and effectiveness, including synchronization, user engagement, and personalization. Our synthesis followed the methodological framework of [54], facilitating systematic comparisons across study groups. Trends and variations in PIMSs’ outcomes were interpreted through subgroup analyses, accounting for methodological rigor and study design. We also considered sample charac- teristics, including demographic variability (eg, age, fitness level, and population type) and sample size heterogeneity (ranging from n=10 to n=150). Limitations arising from study heterogeneity were explicitly addressed to provide transpar- ency regarding factors affecting generalizability. Hedges g effect sizes and SEs were calculated using the tool by Wilson [55]. Meta-analytic models were con- ducted in R (version 4.5.1) using the metafor package [56], applying a random-effects model with the DerSimonian- Laird estimator for physical activity level, physical exertion, RPE, and affective valence. These outcomes were selected based on the preregistration criterion: “meta-analyses will be performed when at least three studies provide data sufficient for effect size calculation.” For inclusion in the meta-analy- sis, physical activity outcomes analyzed included behaviors such as walking, running, weight training, cycling, house- work, and gardening, while studies focusing on nonphysical activity outcomes (eg, subjective feasibility of PIMSs) were excluded. Six studies (comprising 17 intervention arms) met this criterion, while outcomes with insufficient data were excluded. Heterogeneity was assessed using the I² statistic (relative proportion of variability attributable to heterogeneity), τ² statistic (absolute variance), and Cochran Q statistic (a formal test of homogeneity). To address the dispersion of effects across studies, the prediction interval was calculated, as it provides insights into the range of effects expected in future comparable studies, beyond the mean effect size [57]. A sensitivity analysis was conducted to evaluate publication bias by examining the relationship between SEs and effect size estimates. Following the studies by Sterne and Egger [58] and Sterne and Harbord [59], funnel plots were produced to assess asymmetry, while forest plots were used to summarize the data. Data and syntax files for these analyses are available in Multimedia Appendix 2. An exploratory meta-regression analysis was conducted to investigate potential moderators contributing to variabil- ity in the effectiveness of PIMSs on physical activity level, affect, RPE, and physical exertion. Candidate moderators were selected based on their theoretical relevance to physical activity and exercise research: study size, participant age, exercise intensity, and music tempo. Music tempo was categorized into tempo ranges to standardize data across studies with differing methodologies, reflecting its established influence on motivational and psychophysical responses [60]. Exercise intensity was classified using MET guidelines to enhance comparability [61]. Participant age and study size were included to address population-level and methodological variability, respectively. Due to the small number of studies included in the meta-analyses, the meta-regression encom- passed all outcomes of interest, with a focus on generating hypotheses for future research. Specifically, within the meta-regression analysis, music tempo was categorized into 3 distinct groups based on beats per minute (BPM): slow (60‐90 BPM), coded as 1; medium (91‐130 BPM), coded as 2; and fast (131+ BPM), coded as 3. When studies reported variable tempos, the average BPM or the dominant tempo range was used for classifica- tion. Exercise intensity was categorized using a 3-level scale aligned with MET guidelines [61]: low (<3 METs), coded as 1; moderate (3‐6.9 METs), coded as 2; and high (≥7 METs), coded as 3. For studies that did not explicitly report METs, intensity was inferred from descriptions of the exercise type or target heart rate zones. Participant age was handled as follows: for studies reporting mean age directly, the provided value was used. In studies reporting age ranges, the midpoint of the range was used as an estimate. If group-specific mean ages were available, a weighted average was calculated based on group sizes to derive an overall mean age for the study. Results Study Selection All records were excluded by ASReview [27,28]. A total of 523 papers were identified through the initial strategic search using the specified keywords. During the screening process, 3 papers were excluded as duplicates, while 4 additional papers were identified as ineligible based on the inclusion criteria. After screening titles and abstracts, 494 papers were excluded for not meeting the inclusion criteria. Subsequently, 23 full-text papers were assessed for eligibility. Of these, 5 papers were excluded [21,62-65] because they did not evaluate the desired effect or outcome. In total, 18 papers were eligible to be included in this review study (Figure 1). JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 9 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 Figure 1. PRISMA information flow describing the screening process. PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analy- ses. Study Characteristics The study characteristics (Table 1) encompass a diverse range of studies conducted across various countries, including Canada, Spain, Germany, Taiwan, Singapore, Denmark, Finland, Belgium, Switzerland, the Czech Republic, the Netherlands, Norway, and locations not specified. These studies, conducted between 2010 and 2024, provide a broad age range among participants, with some studies focusing on specific groups such as older adults, patients with cardio- vascular disease, students, nonathletes, and office workers. The PIMSs used in these studies vary in their design and objectives, ranging from personalized music audio playlists [30,34,42] to interactive music systems linked to fitness devices [22,35]. These systems are used in different set- tings and for various purposes, ranging from synchronizing movement during physical activity and exercise to enhancing the experience of physical activity and exercise. Reported Outcome Measures A variety of outcome measures were reported across studies to explore the effects of PIMSs on physical activity and exercise-related behaviors. The outcome measures included assessments of physical activity levels, such as accelerometer- based metrics and adherence to specific heart rate zones, as well as psychological and perceptual outcomes such as mood (measured through tools such as the MDMQ and FS) and intrinsic motivation (measured via the Intrinsic Motivation Inventory, IMI). The RPE was frequently captured using the Borg Category-Ratio 10 Scale [51,52]. Table 3 presents this information. Further information on these outcome measure- ments can be found in Section S2 in Multimedia Appendix 1. Studies used diverse technologies and protocols to assess PIMSs’ effects on physical activity and exercise behaviors. Reported technologies included accelerometers, heart rate monitors, and systems such as JYMMiN, which integrate real-time musical feedback with gym equipment. Analyti- cal methods, such as ANOVA and multivariate analysis of variance, were used to evaluate outcomes, with specific systems adapting music based on cadence, heart rate, and intensity. Table 4 presents a detailed overview of these technologies and protocols. JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 10 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 Table 3. Outcome measures as reported in the studies. Outcome measure and measurement method Study Physical activity level Mean weekly minutes of physical activity measured using a triaxial accelerometer. [29] Duration of exercise until exhaustion, timed with a stopwatch. [22] Compliance with exercise regime by monitoring adherence to target heart rate zones during cycling. [12] Affective valence Feeling Scale based on Russell circumplex model of affect. [32] Multidimensional Mood Questionnaire, evaluating “good versus bad” mood dimensions during acute physical exercise. [22,35] Interest or enjoyment subscale of the Intrinsic Motivation Inventory for intrinsic motivation. [12,66,67] Ratings of perceived exertion Borg Category-Ratio 10 Scale, with ratings of perceived exertion collected at specific time intervals during exercise. [12,22,32] Physical exertion Heart rate measured using a Polar Verity Sense (Polar Electro) device based on photoplethysmography. [32] Pace measured via swings per minute using smartphone accelerometer data. [36] Heart rate measured using a Polar T61 (Polar Electro) heart rate belt. [12] Table 4. Technologies and analysis protocols of PIMSsa. Type of PIMSs PIMS description Data analysis protocol Reference Accelerometer Music synchronization with step cadence N/Ab [36] Heart rate monitor Music tempo adjustments based on physiological data ANOVA and MANOVAc [32] JYMMiN Music feedback system ANOVA, MANOVA, and Wilcoxon Signed-Ranks Test [22,35] Magnet and heart rate sensors Magnet sensors detect the ratings of perceived exertion, paired with heart rate, to optimize cycling rhythms N/A [37] moBeat Music feedback system ANOVA [12] Musical sonification systems Converts movement data into sound to enhance engagement and differentiate activity patterns One proportion z-test [39] Musical sonification (custom pedals) Pedals with load sensors and a microcontroller adjust musical feedback ANOVA, Friedman test, and pairwise comparisons [38] Personal activity monitor N/A Generalized linear modeling [29] Three-axis accelerometer (smartphone) Adjusts music tempo to synchronize with the user’s pace using swings per minute N/A [36] Triaxial accelerometer Uses accelerometer and heart rate to adjust music tempo for maintaining the target heart rate during cardio training N/A [33] aPIMS: Personalized Interactive Music System. bN/A: not applicable. cMANOVA: multivariate analysis of variance. Risk of Bias in Studies Following the assessment of the study quality using the JBI critical appraisal checklist tools, the nine criteria were adapted to the 5 risk-of-bias domains found in the R package for risk-of-bias assessments (robvis) in the study by McGuin- ness and Higgins [68]. This assessment tool tests the risk of bias resulting from the randomization process (domain [D1]), deviations from intended intervention (D2), missing outcome data (D3), measurement of the outcome (D4), and selection of the reported result (D5). Each domain is assessed with a judgment scale indicating a high risk of bias (red cross), some concerns (yellow circle), low risk of bias (green plus), and no information (blue question mark; cf Figure 2). We included all 18 studies in the review regardless of their overall risk of bias rating (see Figure 2, column “overall”). The overall risk of bias rating for each study was assigned conservatively, reflecting the highest risk level present across any of the 5 domains (D1-D5). For example, if 1 domain was judged to have a high risk of bias, the overall rating for that study was classified as high risk. Of the 18 studies, 1 randomized experimental design study [29] was rated for JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 11 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 low risk of bias. Seven studies received a moderate (some concerns) rating of risk of bias, and 10 were rated for a high risk of bias. Figure 2. Evaluation of risk of bias in the included studies, categorized across 5 domains from D1 to D5 (cf [68]). An overall bias risk assessment for each study is also provided, conservatively summarizing the findings across all 5 domains [12,22,29-44]. D: domain. PIMSs Used in Experimental Studies PIMSs were explored in experimental studies for their influence on physical, psychophysical, and affective exercise–related outcomes. Several studies focused on synchronization and auditory-motor coupling. Moens et al [41] examined beat synchronization using the D-Jogger adaptive music player. They found that initiating music in phase synchrony significantly enhanced consistent sensorimo- tor patterns, while strategies relying on tempo adjustments alone were less effective. Maes et al [38] provided detailed analyses of synchronization strength using SoundBike (Ghent University), where musical sonification significantly increased pedal cadence synchronization with external music. Similarly, Jun et al [36] found significant increases in step frequency (swings per minute) when music tempo aligned with user pace, enhancing consistency and efficiency of the activity. Rehfeld et al [22] and Fritz et al [35] reported on the JYMMiN system’s role in improving mood and exercise JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 12 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 duration. Notably, Fritz et al [35] observed mood enhance- ments in younger adults, while [22] noted prolonged exercise durations in older adult participants despite no significant mood changes. This may be potentially due to age-related differences in energy pacing. Rosseland [44] explored a tempo-responsive system for Alzheimer patients, observing improved synchronization and engagement. Additionally, van der Vlist et al [12] reported the moBeat system maintained exercise compliance while enhancing intrinsic motivation and attentional dissociation from discomfort. Sample sizes varied (N=10 to N=150), with participants aged 18‐79 years across diverse demographics. Detailed descriptions of PIMSs used in these studies can be found in Table S1 in Multimedia Appendix 3. PIMSs Used in Proof of Concept and User Testing Studies Proof-of-concept and user-testing studies used PIMSs to adapt music or audio feedback based on real-time physi- cal activity and exercise-related data (eg, heart rate, oxy- gen saturation, and galvanic skin response), with a focus on music recommendation systems and synchronization features. Álvarez et al [30] tested DJ-Running (University of Zaragoza), which integrates environmental (GPS) and galvanic skin response data to provide personalized music recommendations using algorithms such as artificial neural networks. Ospina-Bohórquez et al [42] developed a context- aware recommender system using smartphone sensors to adjust music based on exercise intensity, providing evidence for preliminary efficacy in low-concentration activities (eg, low-to-moderate intensity activities that require minimal concentration, such as walking). Two synchronization-based systems were included: [33] music-assisted run trainer, which adjusts music tempo to heart rate or step frequency; and [31] music feedback exercise system, which synchronizes music with movement intensity through advanced audio processing. For example, as exercise intensity increases, additional layers of musical elements such as rhythm guitar, bass, or drums are progressively added to the audio track. Mendoza et al [39] introduced musi- cal sonification, converting movement data into music for users to identify different physical activity patterns. Macule- wicz and Serafin [37] examined ecological soundscapes to influence cycling behavior. Soundscapes were, for example, dynamically altered based on users’ cycling speed and heart rate. Moens et al [40] reported optimal movement entrainment at ~120 BPM using D-Jogger but noted disruptions during song transitions. The reinforcement learning–based system by Fang et al [34] found improved user satisfaction and fewer track rejections, while Rosseland [44] found tempo-respon- sive music systems beneficial for older adults with Alzheimer disease. Details on these systems can be found in Table S2 in Multimedia Appendix 3. Meta-Analyses A single overall meta-analysis of the studies was not achievable due to heterogeneity across datasets and outcomes [69]. Instead, the outcomes were reported separately based on their focus. The reported outcomes distinguished between (1) physical activity levels, (2) physical exertion, (3) RPE, and (4) affective valence. Results for Physical Activity Level The overall effect size is 0.49 with a 95% CI of 0.07 to 0.91, and a P value of .02 (k=4, n=76). This indicates that the results are statistically significant, supporting the effective- ness of PIMSs in improving outcomes relating to physical activity level (Figure 3). The random-effects model indicates low heterogeneity (Q=1.65, P=.65, I²=0%, τ2=0.00) between the studies, suggesting it to be negligible. The calculated 95% prediction interval for the true effect size is 0.07 to 0.91, indicating that while the average effect is positive, the range of potential true effects across future studies could include larger positive outcomes. Figure 3. Forest plot of effect sizes for physical activity level outcomes associated with PIMSs [12,22,29]. PIMS: Personalized Interactive Music System; RAS: rhythmic auditory stimulation. JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 13 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 Results for Physical Exertion The overall effect size is 0.78 with a 95% CI of −0.55 to 2.11, and a P value of .25 (k=5, n=142), indicating that the results are not statistically significant and do not support the effectiveness of PIMSs in improving physical exertion outcomes (Figure 4). The random-effects model indicates high heterogeneity (Q=46.96, P≤.001, I²=91%, τ2=2.08) between the studies. The calculated 95% prediction interval for the true effect size is −2.34 to 3.90, indicating the potential for considerable variation in the effects of PIMSs on physical exertion across future studies. Figure 4. Forest plot of effect sizes for physical exertion outcomes associated with PIMSs [12,32,36]. BPM: beats per minute; PIMS: Personalized Interactive Music System. Results for RPE The overall effect size is 0.72 with a 95% CI of −0.13 to 1.58, and a P value of .10 (k=3, n=77), indicating that the results are not statistically significant and do not conclu- sively support the effectiveness of PIMSs in improving RPE outcomes (Figure 5). The random-effects model indicates substantial heterogeneity (Q=10.24, P=.01, I²=80%, τ2=0.45) between the studies. The calculated 95% prediction inter- val for the true effect size is −0.85 to 2.29, reflecting the significant variability in potential outcomes across future studies. Figure 5. Forest plot of effect sizes for RPE outcomes associated with PIMSs [12,22,32]. PIMS: Personalized Interactive Music System; RPE: ratings of perceived exertion. Results for Affective Valence The overall effect size is 1.65 with a 95% CI of 0.35 to 2.96, and a P value of .01 (k=4, n=122), indicat- ing that the results are statistically significant and thus consistent with the effectiveness of PIMSs in improving affective valence outcomes (Figure 6). The random-effects model indicates substantial heterogeneity (Q=36.69, P<.001, I²=92%, τ2=1.59) between the studies. The calculated 95% prediction interval for the true effect size is −1.14 to 4.44, highlighting significant variability in potential outcomes across future studies. JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 14 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 Figure 6. Forest plot of effect sizes for affective valence outcomes associated with PIMSs [12,22,32,35]. PIMS: Personalized Interactive Music System. Meta-Regression Analysis Heterogeneity was identified in the meta-analyses, prompt- ing the use of meta-regression analysis to explore potential moderators of effect sizes. Music tempi showed a statisti- cally significant positive association with effect sizes (β=.62, SE=0.29, P=.031), suggesting that faster tempi may have a significant effect across the outcomes of interest. None of the other predictors, including participant age, exercise intensity, or sample size, demonstrated a significant effect on effect sizes (Table 5 and Figure 7). The overall meta-regression model was not statistically significant, QM(4)=7.03, P=.135, and a substantial portion of heterogeneity remained unex- plained, QE(11)=76.78, P<.001, I2=85.67%, τ2=0.92. This indicates that other, unexplored factors likely contribute to the variability in outcomes. Given the inclusion of all outcomes of interest in this analysis, the potential for residual variability and unaccounted-for heterogeneity is high. Table 5. A summary of the meta-regression analysis. Predictor Estimate SE z P value 95% CI Intercept 1.194 2.965 0.402 .69 −4.619 to 7.006 Age −0.035 0.040 −0.878 .38 −0.114 to 0.043 Music tempo 0.617 0.286 2.155 .031 0.056 to 1.178 Exercise intensity 0.212 1.262 0.168 .87 −2.262 to 2.686 Sample size −0.025 0.062 −0.402 .69 −0.147 to 0.097 JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 15 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 Figure 7. Distribution of study-level effect sizes across music tempo categories. Violin plots illustrate the density and spread of Hedges g values for medium (91‐130 BPM) and fast (131+ BPM) tempo groups. Dots represent individual study estimates; diamonds and error bars indicate group means and 95% CIs, respectively. BPM: beats per minute. Publication Bias Analysis (Egger Test) Egger test [70] indicated nonsignificant asymmetry for physical activity level (z=−0.968, P=.333), significant asymmetry for physical exertion (z=2.927, P=.003), nonsignificant asymmetry for RPE (z=0.832, P=.405), and significant asymmetry for affective valence (z=4.961, P<.001; Figure 8). Due to potential publication bias, the summary effect sizes for physical exertion and affective valence outcomes may thus be slightly inflated. JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 16 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 Figure 8. Funnel plots for (A) physical activity level, (B) physical exertion, (C) RPE, and (D) affective valence. RPE: ratings of perceived exertion. Discussion Principal Findings This review aimed to systematically evaluate the effective- ness of PIMSs across physical activity levels, physiological outcomes (eg, heart rate and step frequency), psychophysical outcomes (eg, the RPE), and affective valence in relation to physical activity and exercise behaviors. A central focus was an exploratory meta-analysis of PIMSs across these outcome domains. The exploratory meta-analysis revealed that PIMSs demonstrate favorable effects on physical activity levels and affective valence, with effect size estimates surpassing those of general music listening [6]. However, the certainty of evidence is limited by methodological inconsistencies, a moderate to high risk of bias, and the limited number of published studies eligible for meta-analyses. Importantly, no significant effects were observed for RPE or measured physical exertion. This reflects variability in the psychophysi- cal outcomes associated with interventions using PIMSs. When examining the findings of individual studies separately, they offer preliminary evidence that PIMSs may improve physical, psychophysical, and affective outcomes related to physical activity and exercise. For example [22], observed longer exercise durations during sessions using JYMMiN compared to routines with passive music listening, without significant increases in perceived exertion. Simi- larly, Alter et al [29] reported increased weekly physi- cal activity volumes among cardiovascular disease patients using personalized rhythmic auditory stimulation–enhanced playlists. Additionally, Ren et al [43] provided qualita- tive evidence suggesting that PIMSs may prompt physical activity, such as reducing sitting time in office settings. However, interpreting these findings is challenging due to methodological limitations and variability in population characteristics. Some studies focus on clinical populations, such as cardiovascular disease patients [29], while oth- ers target healthy younger adults [33] or older adult par- ticipants [22]. Several studies lack demographic details entirely, further complicating the assessment of population- specific efficacy. Sample sizes also vary widely, from single participants [31] to larger groups (N=36) [42]. Heterogeneity in PIMSs’ Outcomes: Methodological Influences The wide prediction intervals observed across outcome domains reflect the substantial heterogeneity in PIMSs’ effects. For example, prediction intervals for physical activity levels and affective valence highlight significant variability JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 17 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 in potential effect sizes. This suggests that while PIMSs may provide positive average effects, individual study outcomes could range from substantial benefits to negligible or even negative impacts. Similarly, the prediction intervals for RPE and physical exertion emphasize uncertainty surrounding these psychophysical outcomes, pointing to inconsistencies in measurement and intervention design. Specifically, variations in study methodologies and control group conditions contribute significantly to this heterogene- ity. Some studies used passive music or other auditory stimuli as controls, while others used no-music conditions. This negatively affects comparability. Well-powered randomized designs, such as that by Alter et al [29] produced robust findings, whereas smaller studies, such as that by Rehfeld et al [22] yielded nonsignificant results, pointing to the influence of study design and statistical power. Additionally, short intervention durations and small sample sizes [37,39] constrain the generalizability of findings. The absence of standardized metrics and protocols across studies further hinders the ability to synthesize outcomes and develop systematic guidelines for PIMS interventions. To alleviate this, future research should adopt standardized protocols and outcome measures. This could be achieved via a music selection and delivery protocol, ensuring uniformity through a predefined library of music tracks categorized by tempo and intensity, delivered via standardized systems (eg, wireless headphones at consistent volumes). Validated tools such as the Borg RPE and the FS for measuring affective valence, administered at fixed intervals, may enhance comparability. Feasibility of PIMSs on Physical Activity Levels and Affective Outcomes Despite methodological inconsistencies, our findings suggest that PIMSs may have a positive influence on physical activity levels. Studies in this cluster were rated as having low [29] to moderate [12,22] risk of bias, with both the studies by Alter et al [29] and Rehfeld et al [22] focusing on older adult populations. Positive effects include increased exercise duration (~66 seconds) [22] and overall weekly physical activity (~105.4 additional minutes per week on average) [29]. However, van der Vlist et al [12] found no signifi- cant impact of PIMSs on physical activity levels. The low heterogeneity in this cluster indicates consistent findings despite variations in study design and participant populations. This is promising and calls for further investigation. Our results align with that of Clark et al [6], who noted that music listening, when combined with physical activity, enhances exercise outcomes in older adults. Both Alter et al [29] and van der Vlist et al [12] used synchronization strategies—rhythmic auditory stimulation and auditory-motor coupling, respectively—consistent with frameworks by Bood et al [8] and Clark et al [71] that link synchronized music to improved physical activity and exercise performance. However, the exploratory nature of the meta-analysis and the small number of studies limit the potential generalizability of these findings. Further research with diverse populations and robust methodologies is required to confirm whether PIMSs are effective adjuncts for increasing physical activity levels. For affective valence, the large effect size estimate suggests PIMSs contribute to elevated affective experiences during physical activity and exercise [12,22,32,35]. However, this finding is strongly influenced by van der Vlist et al [12], whose notably high effect size estimate substantially raised the overall meta-analytic effect size estimate. In contrast, smaller effects observed in other studies [22,35] reduced the precision and generalizability of the overall meta-analytic finding. The differences in these outcomes likely reflect variations in music selection methods: researcher-selected music in the study by van der Vlist et al [12] prompted synchronization and enjoyment (“fun and enjoyment” ratings via IMI), while self-selected music [32] and device-gen- erated feedback [22,35] influenced affective outcomes in distinct ways. In the study by van der Vlist [12], researcher- selected music facilitated synchronization, while Chen et al [32] used self-selected music based on participants’ individ- ual preferences. Rehfeld et al [22] and Fritz et al [35] used device-generated musical feedback, where participants’ movements influenced the music. These differences suggest that PIMSs may enhance affective valence outcomes during physical activity and exercise through both self-selected and researcher-selected music, with evidence of positive effects for music tailored to individual preferences (aligning with prior research by Terry et al [7] and Khalfa et al [11]) as well as for standardized, researcher-selected stimuli. Curiously, van der Vlist et al [12] reported no signifi- cant benefits for RPE, despite using auditory-motor cou- pling strategies. This discrepancy may find alignment with the Dual-Mode Theory, as even though music can enhance automatic synchronization and facilitate improved physical performance, it does not always mitigate RPE if reflective processes (eg, cognitive appraisal of effort) are less engaged [13]. The substantial heterogeneity within the affective valence cluster, driven by variability in musical strategies, participant demographics, and inconsistent measurement tools (eg, MDMQ, IMI, and FS), further supports ART’s asser- tion that individual and contextual factors critically shape affective outcomes during exercise. All studies in the affective valence cluster were deemed to have a moderate risk of bias. Furthermore, the reliance on measurement scales without strong theoretical grounding, as noted in the study by van der Vlist et al [12], suggests the need for alignment with validated frameworks such as ART. For instance, the FS used by Chen et al [32] directly measures the pleasure-displeasure dimensions central to ART, aligning with validated frameworks in physical activity and exercise contexts [72]. The FS provides a theoretically robust and context-specific assessment of affective responses, captur- ing the transient emotional states during exercise that ART posits are critical for shaping future behavioral intentions. These findings tentatively indicate that these PIMSs leverage momentary affective responses to improve exercise experien- ces [6,7,17]. In sum, findings across the physical activity and affective valence meta-analytic clusters suggest PIMSs may support affect augmentation during physical activity, highlighting their potential to enhance both physical activity levels and affective outcomes [5,17]. JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 18 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 PIMSs’ Tempo Adjustments and Synchronization in Physical Activity and Exercise Outcomes The identification of faster music tempi as a statistically significant moderator in the meta-regression aligns with evidence supporting the role of synchronization strength and auditory-motor coupling in enhancing exercise outcomes [8,60]. For instance, faster tempi provide consistent rhyth- mic cues that facilitate the alignment of motor actions with auditory stimuli. This can optimize auditory-motor coupling [8-10], which, in turn, enables predictive synchronization to reduce RPE [7]. For example, Chen et al [32] reported that real-time tempo adjustments based on heart rate signifi- cantly reduced RPE and improved affective responses. This indicates that synchronized music facilitated participants’ dissociation from internal sensory signals and promoted enjoyment during exercise [7]. Limitations and Future Directions This review presents the first systematic exploration of PIMSs exclusively within physical, psychophysical, and affective domains of physical activity and exercise. While it provides valuable insights, several limitations must be acknowledged. A significant proportion of the included studies (14 of 18) primarily assessed the feasibility of PIMSs, with few investigating direct outcomes related to physical activity or exercise. Many experimental studies were limited by short durations, small sample sizes, and insufficiently rigorous methodologies. Similarly, proof-of-concept and user-testing studies largely focused on system feasibility rather than assessing objective psychophysiological outcomes. Conse- quently, the high risk of bias in 10 studies underscores the overall low quality of evidence. Additionally, the small number of eligible studies precluded sensitivity analyses, which further emphasizes the preliminary nature of this review’s findings. Few studies identified physical activity as a primary outcome, often relegating it to secondary importance. Objective assessments of physical activity—such as measures of frequency, intensity, and duration—were notably absent, making it difficult to draw robust conclusions or compare results across studies. Standardizing methods for quantifying physical activity would enhance future research by enabling more meaningful cross-study comparisons. Furthermore, the methodology used in this study was limited by substantial heterogeneity across studies. This prevents a unified meta-analysis and necessitates the reporting of separate outcomes. Variability in study designs, participant demographics, and measurement tools contrib- uted to unexplained heterogeneity, while the small num- ber of studies precluded sensitivity analyses. These factors, combined with the exploratory nature of the meta-analysis, point to the need for standardized methodologies and rigorous reporting in future research. Additional limitations include the possibility of publication and retrieval bias, as only Eng- lish-language studies from selected databases were included. Furthermore, although screening and data extraction were independently conducted by 2 reviewers, the use of automa- ted tools and subjective judgment may have introduced bias. To address these limitations, future research should prioritize larger, randomized controlled trials with diverse populations and longer intervention periods. Longitudinal studies are particularly needed to evaluate the sustained impact of PIMSs on physical activity and exercise. Addi- tionally, investigating the mechanisms underlying individ- ual variability in PIMSs’ responses could optimize these systems for different populations and exercise contexts. This highlights the need for more rigorous research to validate these effects and refine PIMSs’ interventions, particularly through the development of dynamic systems that can adapt tempo in real time to suit diverse user needs and exercise contexts [41,71]. Emerging trends in PIMSs, such as music recommender systems examined by Álvarez et al [30], Fang et al [34], and Ospina-Bohórquez et al [42], highlight the potential for integration with streaming services such as Spotify (Spotify AB). These systems demonstrated promising user feedback [34] and feasibility, suggesting they could serve as a foundation for future hypothesis-driven studies. Incorporat- ing feedback from wearable and smartphone devices offers another avenue for development, allowing PIMSs to adapt based on physical activity and exercise metrics as well as music preferences. Finally, many PIMSs are relatively low-cost interventions (eg, the devices in the study by Alter et al [29] cost approximately US $75 per patient) and could have significant cost-effectiveness implications as part of broader health policy strategies to enhance physical activity and exercise participation at the population level [5]. Conclusions This systematic review provides exploratory evidence that PIMSs may positively impact physical activity levels and affective valence in physical activity and exercise contexts. The meta-analysis revealed moderate effect sizes for physical activity levels and significant but heterogeneously distributed effects for affective valence. However, outcomes for RPE and physical exertion were inconclusive due to high heterogeneity and limited study quality. The findings are constrained by methodological limita- tions, including high risk of bias, small sample sizes, short study durations, and inconsistent measures across studies. Furthermore, the lack of theoretical frameworks for inform- ing PIMSs’ designs and the absence of standardization in quantifying physical activity outcomes limit the generaliza- bility of these findings. PIMSs remain considerably underex- plored, and further research is essential. Overall, PIMSs provide promising potential for enhanc- ing physical activity levels and elevated affective valence, offering engaging physical activity and exercise opportuni- ties for the public at large. With advancements in adap- tive systems capable of real-time tempo adjustments, PIMSs may emerge as effective adjuncts for physical activity and exercise, pending rigorous validation in diverse populations. JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 19 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 Acknowledgments This project has received funding from the Research Council of Finland (346210). Authors’ Contributions AD, TK, GL, and SC conceptualized this study. AD, TK, and GL handled the methodology. AD, TK, IB, P Neto, AM, and WMR investigated. AD, TK, P Nijhuis, IB, P Neto, AM, and WMR curated the data. AD and KK worked on the formal analysis. AD, FK, P Nijhuis, RR, KRA, JM, and NCH wrote the original draft. AD, TK, FK, KK, P Nijhuis, IB, P Neto, AM, WMR, NCH, AA, TR, VA, MH, RSS, JKI, RR, KRA, JM, PT, SS, SC, and GL reviewed and edited the writing. Conflicts of Interest None declared. Checklist 1 PRISMA checklist. PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses. [DOCX File (Microsoft Word File), 24 KB-Checklist 1] Multimedia Appendix 1 Descriptions and outcomes of the PIMSs used across studies in sections. PIMS: Personalized Interactive Music System. [DOCX File (Microsoft Word File), 29 KB-Multimedia Appendix 1] Multimedia Appendix 2 Data and syntax for meta-analysis. [DOCX File (Microsoft Word File), 14 KB-Multimedia Appendix 2] Multimedia Appendix 3 Descriptions and outcomes of the PIMSs used across studies in tables. PIMS: Personalized Interactive Music System. [DOCX File (Microsoft Word File), 34 KB-Multimedia Appendix 3] References 1. Bull FC, Al-Ansari SS, Biddle S, et al. World Health Organization 2020 guidelines on physical activity and sedentary behaviour. Br J Sports Med. Dec 2020;54(24):1451-1462. [doi: 10.1136/bjsports-2020-102955] 2. Posadzki P, Pieper D, Bajpai R, et al. Exercise/physical activity and health outcomes: an overview of Cochrane systematic reviews. BMC Public Health. Nov 16, 2020;20(1):1724. [doi: 10.1186/s12889-020-09855-3] [Medline: 33198717] 3. Caspersen CJ, Powell KE, Christenson GM. Physical activity, exercise, and physical fitness: definitions and distinctions for health-related research. Public Health Rep. 1985;100(2):126-131. [Medline: 3920711] 4. Chase JAD. Systematic review of physical activity intervention studies after cardiac rehabilitation. J Cardiovasc Nurs. 2011;26(5):351-358. [doi: 10.1097/JCN.0b013e3182049f00] [Medline: 21263342] 5. Moy ML, Teylan M, Weston NA, Gagnon DR, Garshick E. Daily step count predicts acute exacerbations in a US cohort with COPD. PLoS ONE. 2013;8(4):e60400. [doi: 10.1371/journal.pone.0060400] [Medline: 23593211] 6. Clark IN, Taylor NF, Peiris CL. Music listening interventions for physical activity: a systematic review and meta- analysis of randomised controlled trials. Disabil Rehabil. Jan 2024;46(1):13-20. [doi: 10.1080/09638288.2022.2155715] [Medline: 36523132] 7. Terry PC, Karageorghis CI, Curran ML, Martin OV, Parsons-Smith RL. Effects of music in exercise and sport: a meta- analytic review. Psychol Bull. Feb 2020;146(2):91-117. [doi: 10.1037/bul0000216] [Medline: 31804098] 8. Bood RJ, Nijssen M, van der Kamp J, Roerdink M. The power of auditory-motor synchronization in sports: enhancing running performance by coupling cadence with the right beats. PLoS ONE. 2013;8(8):e70758. [doi: 10.1371/journal. pone.0070758] [Medline: 23951000] 9. Karageorghis CI, Priest DL. Music in the exercise domain: a review and synthesis (part I). Int Rev Sport Exerc Psychol. Mar 2012;5(1):44-66. [doi: 10.1080/1750984X.2011.631026] [Medline: 22577472] 10. Edworthy J, Waring H. The effects of music tempo and loudness level on treadmill exercise. Ergonomics. Dec 15, 2006;49(15):1597-1610. [doi: 10.1080/00140130600899104] [Medline: 17090506] 11. Khalfa S, Roy M, Rainville P, Dalla Bella S, Peretz I. Role of tempo entrainment in psychophysiological differentiation of happy and sad music? Int J Psychophysiol. Apr 2008;68(1):17-26. [doi: 10.1016/j.ijpsycho.2007.12.001] [Medline: 18234381] JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 20 (page number not for citation purposes) https://jmir.org/api/download?alt_name=humanfactors_v12i1e70372_app1.docx https://jmir.org/api/download?alt_name=humanfactors_v12i1e70372_app1.docx https://jmir.org/api/download?alt_name=humanfactors_v12i1e70372_app2.docx https://jmir.org/api/download?alt_name=humanfactors_v12i1e70372_app2.docx https://jmir.org/api/download?alt_name=humanfactors_v12i1e70372_app3.docx https://jmir.org/api/download?alt_name=humanfactors_v12i1e70372_app3.docx https://jmir.org/api/download?alt_name=humanfactors_v12i1e70372_app4.docx https://jmir.org/api/download?alt_name=humanfactors_v12i1e70372_app4.docx https://doi.org/10.1136/bjsports-2020-102955 https://doi.org/10.1186/s12889-020-09855-3 http://www.ncbi.nlm.nih.gov/pubmed/33198717 http://www.ncbi.nlm.nih.gov/pubmed/3920711 https://doi.org/10.1097/JCN.0b013e3182049f00 http://www.ncbi.nlm.nih.gov/pubmed/21263342 https://doi.org/10.1371/journal.pone.0060400 http://www.ncbi.nlm.nih.gov/pubmed/23593211 https://doi.org/10.1080/09638288.2022.2155715 http://www.ncbi.nlm.nih.gov/pubmed/36523132 https://doi.org/10.1037/bul0000216 http://www.ncbi.nlm.nih.gov/pubmed/31804098 https://doi.org/10.1371/journal.pone.0070758 https://doi.org/10.1371/journal.pone.0070758 http://www.ncbi.nlm.nih.gov/pubmed/23951000 https://doi.org/10.1080/1750984X.2011.631026 http://www.ncbi.nlm.nih.gov/pubmed/22577472 https://doi.org/10.1080/00140130600899104 http://www.ncbi.nlm.nih.gov/pubmed/17090506 https://doi.org/10.1016/j.ijpsycho.2007.12.001 http://www.ncbi.nlm.nih.gov/pubmed/18234381 https://humanfactors.jmir.org/2025/1/e70372 12. van der Vlist B, Bartneck C, Mäueler S. moBeat: using interactive music to guide and motivate users during aerobic exercising. Appl Psychophysiol Biofeedback. Jun 2011;36(2):135-145. [doi: 10.1007/s10484-011-9149-y] [Medline: 21380562] 13. Ekkekakis P. Pleasure and displeasure from the body: perspectives from exercise. Cogn Emot. Mar 2003;17(2):213-239. [doi: 10.1080/02699930302292] [Medline: 29715726] 14. Russell JA. Core affect and the psychological construction of emotion. Psychol Rev. Jan 2003;110(1):145-172. [doi: 10. 1037/0033-295x.110.1.145] [Medline: 12529060] 15. Brand R, Ekkekakis P. Affective–Reflective Theory of physical inactivity and exercise. Ger J Exerc Sport Res. Mar 2018;48(1):48-58. [doi: 10.1007/s12662-017-0477-9] 16. Karageorghis CI. The scientific application of music in exercise and sport: towards a new theoretical model. In: Sport and Exercise Psychology. New York, NY, US: Routledge/Taylor & Francis Group; 2016:276-322. 17. Karageorghis CI. Music‐Related interventions in the exercise domain. In: Handbook of Sport Psychology. Wiley; 2020:929-949. [doi: 10.1002/9781119568124] 18. Agres KR, Schaefer RS, Volk A, et al. Music, computing, and health: a roadmap for the current and future roles of music technology for health care and well-being. Music Sci. Jan 1, 2021;4. [doi: 10.1177/2059204321997709] 19. Danso A. The use of technology in music-based interventions for health and education [Dissertation]. Jyväskylä: University of Jyväskylä; 2023. 20. Marcus BH, Owen N, Forsyth LH, Cavill NA, Fridinger F. Physical activity interventions using mass media, print media, and information technology. Am J Prev Med. Nov 1998;15(4):362-378. [doi: 10.1016/s0749-3797(98)00079-8] [Medline: 9838978] 21. Wijnalda G, Pauws S, Vignoli F, Stuckenschmidt H. A personalized music system for motivation in sport performance. IEEE Pervasive Comput. Jul 2005;4(3):26-32. [doi: 10.1109/MPRV.2005.47] 22. Rehfeld K, Fritz TH, Prinz A, Schneider L, Villringer A, Witte K. Musical feedback system JYMMiN leads to enhanced physical endurance in the elderly-a feasibility study. Front Sports Act Living. 2022;4:915926. [doi: 10.3389/fspor.2022. 915926] [Medline: 36032261] 23. Laranjo L, Ding D, Heleno B, et al. Do smartphone applications and activity trackers increase physical activity in adults? Systematic review, meta-analysis and metaregression. Br J Sports Med. Apr 2021;55(8):422-432. [doi: 10.1136/bjsports- 2020-102892] [Medline: 33355160] 24. Michie S, Richardson M, Johnston M, et al. The behavior change technique taxonomy (v1) of 93 hierarchically clustered techniques: building an international consensus for the reporting of behavior change interventions. Ann Behav Med. Aug 2013;46(1):81-95. [doi: 10.1007/s12160-013-9486-6] [Medline: 23512568] 25. Ziv G, Lidor R. Music, exercise performance, and adherence in clinical populations and in the elderly: a review. J Clin Sport Psychol. Mar 2011;5(1):1-23. [doi: 10.1123/jcsp.5.1.1] 26. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. Mar 29, 2021;372:n71. [doi: 10.1136/bmj.n71] [Medline: 33782057] 27. van de Schoot R, de Bruin J, Schram R, et al. An open source machine learning framework for efficient and transparent systematic reviews. Nat Mach Intell. Feb 2021;3(2):125-133. [doi: 10.1038/s42256-020-00287-7] 28. Ouzzani M, Hammady H, Fedorowicz Z, Elmagarmid A. Rayyan-a web and mobile app for systematic reviews. Syst Rev. Dec 5, 2016;5(1):210. [doi: 10.1186/s13643-016-0384-4] [Medline: 27919275] 29. Alter DA, O’Sullivan M, Oh PI, et al. Synchronized personalized music audio-playlists to improve adherence to physical activity among patients participating in a structured exercise program: a proof-of-principle feasibility study. Sports Med Open. 2015;1(1):23. [doi: 10.1186/s40798-015-0017-9] [Medline: 26284164] 30. Álvarez P, Zarazaga-Soria FJ, Baldassarri S. Mobile music recommendations for runners based on location and emotions: the DJ-Running system. Pervasive Mob Comput. Sep 2020;67:101242. [doi: 10.1016/j.pmcj.2020.101242] 31. Carôt A, Fritz TH, Englert K. Towards a system supporting music feedback exercise in physical tele-rehabilitation. Presented at: 2023 4th International Symposium on the Internet of Sounds; Oct 26-27, 2023:1-7; Pisa, Italy. [doi: 10. 1109/ieeeconf59510.2023.10335285] 32. Chen Y, Chen CC, Tang LC, Chieng WH. Mid-distance running exercise assistance system via iot and exercise-oriented music. Presented at: 2023 IEEE World AI IoT Congress (AIIoT); Jun 7-10, 2023:264-272; Seattle, WA, USA. [doi: 10. 1109/AIIoT58121.2023.10174324] 33. Chen LY, Tung YJ, Jang JSR. MART-music assisted run trainer. Presented at: 15th International Society for Music Information Retrieval Conference; Oct 27-31, 2014; Taipei, Taiwan. 34. Fang J, Grunberg D, Lui S, Wang Y. Development of a music recommendation system for motivating exercise. Presented at: 2017 International Conference on Orange Technologies (ICOT); Dec 8-10, 2018:83-86; Singapore. 2018. [doi: 10.1109/ICOT.2017.8336094] JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 21 (page number not for citation purposes) https://doi.org/10.1007/s10484-011-9149-y http://www.ncbi.nlm.nih.gov/pubmed/21380562 https://doi.org/10.1080/02699930302292 http://www.ncbi.nlm.nih.gov/pubmed/29715726 https://doi.org/10.1037/0033-295x.110.1.145 https://doi.org/10.1037/0033-295x.110.1.145 http://www.ncbi.nlm.nih.gov/pubmed/12529060 https://doi.org/10.1007/s12662-017-0477-9 https://doi.org/10.1002/9781119568124 https://doi.org/10.1177/2059204321997709 https://doi.org/10.1016/s0749-3797(98)00079-8 http://www.ncbi.nlm.nih.gov/pubmed/9838978 https://doi.org/10.1109/MPRV.2005.47 https://doi.org/10.3389/fspor.2022.915926 https://doi.org/10.3389/fspor.2022.915926 http://www.ncbi.nlm.nih.gov/pubmed/36032261 https://doi.org/10.1136/bjsports-2020-102892 https://doi.org/10.1136/bjsports-2020-102892 http://www.ncbi.nlm.nih.gov/pubmed/33355160 https://doi.org/10.1007/s12160-013-9486-6 http://www.ncbi.nlm.nih.gov/pubmed/23512568 https://doi.org/10.1123/jcsp.5.1.1 https://doi.org/10.1136/bmj.n71 http://www.ncbi.nlm.nih.gov/pubmed/33782057 https://doi.org/10.1038/s42256-020-00287-7 https://doi.org/10.1186/s13643-016-0384-4 http://www.ncbi.nlm.nih.gov/pubmed/27919275 https://doi.org/10.1186/s40798-015-0017-9 http://www.ncbi.nlm.nih.gov/pubmed/26284164 https://doi.org/10.1016/j.pmcj.2020.101242 https://doi.org/10.1109/ieeeconf59510.2023.10335285 https://doi.org/10.1109/ieeeconf59510.2023.10335285 https://doi.org/10.1109/AIIoT58121.2023.10174324 https://doi.org/10.1109/AIIoT58121.2023.10174324 https://doi.org/10.1109/ICOT.2017.8336094 https://humanfactors.jmir.org/2025/1/e70372 35. Fritz TH, Halfpaap J, Grahl S, Kirkland A, Villringer A. Musical feedback during exercise machine workout enhances mood. Front Psychol. 2013;4(DEC):921. [doi: 10.3389/fpsyg.2013.00921] [Medline: 24368905] 36. Jun S, Rew J, Hwang E. Runner’s jukebox: a music player for running using pace recognition and music mixing. Presented at: MMEDIA 2015: The Seventh International Conferences on Advances in Multimedia; Apr 19-24, 2015; Barcelona, Spain. 37. Maculewicz J, Serafin S. A stationary bike in augmented audio reality. Presented at: REHAB ’15; Oct 1-2, 2015:164-167; Lisbon Portugal. Oct 2015.[doi: 10.1145/2838944.2838984] 38. Maes PJ, Lorenzoni V, Six J. The SoundBike: musical sonification strategies to enhance cyclists’ spontaneous synchronization to external music. J Multimodal User Interfaces. Sep 2019;13(3):155-166. [doi: 10.1007/s12193-018- 0279-x] 39. Mendoza JI, Danso A, Luck G, Rantalainen T, Palmberg L, Chastin S. Musification of accelerometry data towards raising awareness of physical activity. Presented at: Proceedings of the Conference on Sonification of Health and Environmental Data (SoniHED 2022); Oct 27-28, 2022; Stockholm, Sweden. [doi: 10.5281/ZENODO.7243875] 40. Moens B, van Noorden L, Leman M. D-jogger: syncing music with walking. Presented at: 7th Sound and music computing Conference; Jul 21-24, 2010:451-456; Barcelona. 2010.URL: https://biblio.ugent.be/publication/1070528 [Accessed 2025-08-06] 41. Moens B, Muller C, van Noorden L, et al. Encouraging spontaneous synchronisation with D-Jogger, an adaptive music player that aligns movement and music. PLoS ONE. Dec 2014;9(12):e114234. [doi: 10.1371/journal.pone.0114234] 42. Ospina-Bohórquez A, Gil-González AB, Moreno-García MN, Luis-Reboredo A. Context-Aware Music Recommender System Based on Automatic Detection of the User’s Physical Activity. Springer; 2021:142-151. [doi: 10.1007/978-3- 030-53036-5_15] 43. Ren X, Lu Y, Visser V, Le PDH, van den Burg R. Interaction matters? Exploring interactive music as a reminder to break sedentary office time. Presented at: 34th International Symposium on Automation and Robotics in Construction; Jun 27-30, 2017:1099-1106; Taipei, Taiwan. 2017.[doi: 10.22260/ISARC2017/0151] 44. Rosseland RB. Design and evaluation of an interactive music system for exercise and physical activity with Alzheimer’s patients. SoundEffects. 2016;6(1):4-22. [doi: 10.7146/se.v6i1.24910] 45. Bassett DR, Troiano RP, McClain JJ, Wolff DL. Accelerometer-based physical activity: total volume per day and standardized measures. Med Sci Sports Exerc. Apr 2015;47(4):833-838. [doi: 10.1249/MSS.0000000000000468] [Medline: 25102292] 46. Kujala UM, Pietilä J, Myllymäki T, et al. Physical activity: absolute intensity versus relative-to-fitness-level volumes. Med Sci Sports Exerc. Mar 2017;49(3):474-481. [doi: 10.1249/MSS.0000000000001134] [Medline: 27875497] 47. Hardy CJ, Rejeski WJ. Not what, but how one feels: the measurement of affect during exercise. J Sport Exerc Psychol. Sep 1989;11(3):304-317. [doi: 10.1123/jsep.11.3.304] 48. Steyer R, Schwenkmezger P, Notz P, Eid M. Development of the Multidimensional Mood State Questionnaire (MDBF) Primary Data. Primärdatensatz. ZPID Leibniz Institute for Psychology; 2003. [doi: 10.5160/PSYCHDATA. SRRF91EN15] 49. Ekkekakis P, Petruzzello SJ. Analysis of the affect measurement conundrum in exercise psychology. Psychol Sport Exerc. Oct 2000;1(2):71-88. [doi: 10.1016/S1469-0292(00)00010-8] 50. Nystoriak MA, Bhatnagar A. Cardiovascular effects and benefits of exercise. Front Cardiovasc Med. 2018;5:135. [doi: 10.3389/fcvm.2018.00135] [Medline: 30324108] 51. Borg G. Perceived exertion as an indicator of somatic stress. Scand J Rehabil Med. 1970;2(2):92-98. [doi: 10.2340/ 1650197719702239298] 52. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc. 1982;14(5):377-381. [Medline: 7154893] 53. Grammatopoulos T, Hunter JWS, Munn Z, Stone JC, Barker TH. Reporting quality and risk of bias in JBI systematic reviews evaluating the effectiveness of interventions: a methodological review protocol. JBI Evidence Synthesis. Jan 2023;21(3):584-591. [doi: 10.11124/JBIES-22-00317] 54. Campbell M, McKenzie JE, Sowden A, et al. Synthesis without meta-analysis (SWiM) in systematic reviews: reporting guideline. BMJ. Jan 16, 2020;368:l6890. [doi: 10.1136/bmj.l6890] [Medline: 31948937] 55. Wilson D. Practical meta-analysis effect size calculator. Campbell Collaboration. 2023. URL: https:// campbellcollaboration.org/research-resources/effect-size-calculator.html [Accessed 2025-07-30] 56. Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Soft. 2010;36(3):1-48. [doi: 10.18637/ jss.v036.i03] 57. Borenstein M. How to understand and report heterogeneity in a meta-analysis: the difference between I-squared and prediction intervals. Integr Med Res. Dec 2023;12(4):101014. [doi: 10.1016/j.imr.2023.101014] [Medline: 38938910] JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 22 (page number not for citation purposes) https://doi.org/10.3389/fpsyg.2013.00921 http://www.ncbi.nlm.nih.gov/pubmed/24368905 https://doi.org/10.1145/2838944.2838984 https://doi.org/10.1007/s12193-018-0279-x https://doi.org/10.1007/s12193-018-0279-x https://doi.org/10.5281/ZENODO.7243875 https://biblio.ugent.be/publication/1070528 https://doi.org/10.1371/journal.pone.0114234 https://doi.org/10.1007/978-3-030-53036-5_15 https://doi.org/10.1007/978-3-030-53036-5_15 https://doi.org/10.22260/ISARC2017/0151 https://doi.org/10.7146/se.v6i1.24910 https://doi.org/10.1249/MSS.0000000000000468 http://www.ncbi.nlm.nih.gov/pubmed/25102292 https://doi.org/10.1249/MSS.0000000000001134 http://www.ncbi.nlm.nih.gov/pubmed/27875497 https://doi.org/10.1123/jsep.11.3.304 https://doi.org/10.5160/PSYCHDATA.SRRF91EN15 https://doi.org/10.5160/PSYCHDATA.SRRF91EN15 https://doi.org/10.1016/S1469-0292(00)00010-8 https://doi.org/10.3389/fcvm.2018.00135 http://www.ncbi.nlm.nih.gov/pubmed/30324108 https://doi.org/10.2340/1650197719702239298 https://doi.org/10.2340/1650197719702239298 http://www.ncbi.nlm.nih.gov/pubmed/7154893 https://doi.org/10.11124/JBIES-22-00317 https://doi.org/10.1136/bmj.l6890 http://www.ncbi.nlm.nih.gov/pubmed/31948937 https://campbellcollaboration.org/research-resources/effect-size-calculator.html https://campbellcollaboration.org/research-resources/effect-size-calculator.html https://doi.org/10.18637/jss.v036.i03 https://doi.org/10.18637/jss.v036.i03 https://doi.org/10.1016/j.imr.2023.101014 http://www.ncbi.nlm.nih.gov/pubmed/38938910 https://humanfactors.jmir.org/2025/1/e70372 58. Sterne JAC, Egger M. Funnel plots for detecting bias in meta-analysis. J Clin Epidemiol. Oct 2001;54(10):1046-1055. [doi: 10.1016/S0895-4356(01)00377-8] [Medline: 11576817] 59. Sterne JAC, Harbord RM. Funnel plots in meta-analysis. Stata J. Jun 2004;4(2):127-141. [doi: 10.1177/ 1536867X0400400204] 60. Elliott D. Music during exercise: does tempo influence psychophysical responses? Philica; 2007. 61. Jetté M, Sidney K, Blümchen G. Metabolic equivalents (METS) in exercise testing, exercise prescription, and evaluation of functional capacity. Clin Cardiol. Aug 1990;13(8):555-565. [doi: 10.1002/clc.4960130809] [Medline: 2204507] 62. Elliott GT, Tomlinson B. Presented at: CHI EA ’06: CHI ’06 Extended Abstracts on Human Factors in Computing Systems; Apr 22-27, 2006:736-741; Montréal Québec Canada. [doi: 10.1145/1125451.1125599] 63. Gallo D, Shreepriya S, Willamowski J. RunAhead: exploring head scanning based navigation for runners. Presented at: CHI ’20: Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems; Apr 25-30, 2020:1-13; Honolulu HI USA. [doi: 10.1145/3313831.3376828] 64. Antón AM, Rodríguez BR. Runtastic PRO app: an excellent all-rounder for logging fitness. Br J Sports Med. Jun 2016;50(11):705-706. [doi: 10.1136/bjsports-2015-094940] [Medline: 26294442] 65. Nymoen K, Song S, Hafting Y, Torresen J. Funky sole music: gait recognition and adaptive mapping. Presented at: Proceedings of the International Conference on New Interfaces for Musical Expression; Jun 30 to Jul 3, 2014; Goldsmiths, University of London, UK. [doi: 10.5281/ZENODO.1178895] 66. Ryan RM, Mims V, Koestner R. Relation of reward contingency and interpersonal context to intrinsic motivation: a review and test using cognitive evaluation theory. J Pers Soc Psychol. Oct 1983;45(4):736-750. [doi: 10.1037/0022- 3514.45.4.736] 67. McAuley E, Duncan T, Tammen VV. Psychometric properties of the Intrinsic Motivation Inventory in a competitive sport setting: a confirmatory factor analysis. Res Q Exerc Sport. Mar 1989;60(1):48-58. [doi: 10.1080/02701367.1989. 10607413] [Medline: 2489825] 68. McGuinness LA, Higgins JPT. Risk-of-bias VISualization (robvis): an R package and Shiny web app for visualizing risk-of-bias assessments. Res Synth Methods. Jan 2021;12(1):55-61. [doi: 10.1002/jrsm.1411] [Medline: 32336025] 69. Cumpston M, Li T, Page MJ, et al. Updated guidance for trusted systematic reviews: a new edition of the Cochrane Handbook for Systematic Reviews of Interventions. Cochrane Database Syst Rev. Oct 3, 2019;10(10):ED000142. [doi: 10.1002/14651858.ED000142] [Medline: 31643080] 70. Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. Sep 13, 1997;315(7109):629-634. [doi: 10.1136/bmj.315.7109.629] [Medline: 9310563] 71. Clark IN, Baker FA, Taylor NF. The modulating effects of music listening on health-related exercise and physical activity in adults: a systematic review and narrative synthesis. Nord J Music Ther. Jan 2, 2016;25(1):76-104. [doi: 10. 1080/08098131.2015.1008558] 72. Bastos V, Rodrigues F, Davis P, Teixeira DS. Assessing affective valence and activation in resistance training with the feeling scale and the felt arousal scale: a systematic review. PLOS ONE. 2023;18(11):e0294529. [doi: 10.1371/journal. pone.0294529] [Medline: 37972201] Abbreviations ART: Affective-Reflective Theory BPM: beats per minute D: domain FS: Feeling Scale IMI: Intrinsic Motivation Inventory JBI: Joanna Briggs Institute MDMQ: Multidimensional Mood Questionnaire MET: metabolic equivalent of task PIMS: Personalized Interactive Music System RPE: ratings of perceived exertion Edited by Andre Kushniruk; peer-reviewed by Alexander Carot, Byron Lai, chandrashekar br, Yonggang Zhang; submitted 20.12.2024; final revised version received 28.05.2025; accepted 26.06.2025; published 08.09.2025 Please cite as: Danso A, Kekäläinen T, Koehler F, Knittle K, Nijhuis P, Burunat I, Neto P, Mavrolampados A, Randall WM, Hansen NC, Ansani A, Rantalainen T, Alluri V, Hartmann M, Schaefer RS, Ihalainen JK, Rousi R, Agres KR, MacRitchie J, Toiviainen P, Saarikallio S, Chastin S, Luck G JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 23 (page number not for citation purposes) https://doi.org/10.1016/S0895-4356(01)00377-8 http://www.ncbi.nlm.nih.gov/pubmed/11576817 https://doi.org/10.1177/1536867X0400400204 https://doi.org/10.1177/1536867X0400400204 https://doi.org/10.1002/clc.4960130809 http://www.ncbi.nlm.nih.gov/pubmed/2204507 https://doi.org/10.1145/1125451.1125599 https://doi.org/10.1145/3313831.3376828 https://doi.org/10.1136/bjsports-2015-094940 http://www.ncbi.nlm.nih.gov/pubmed/26294442 https://doi.org/10.5281/ZENODO.1178895 https://doi.org/10.1037/0022-3514.45.4.736 https://doi.org/10.1037/0022-3514.45.4.736 https://doi.org/10.1080/02701367.1989.10607413 https://doi.org/10.1080/02701367.1989.10607413 http://www.ncbi.nlm.nih.gov/pubmed/2489825 https://doi.org/10.1002/jrsm.1411 http://www.ncbi.nlm.nih.gov/pubmed/32336025 https://doi.org/10.1002/14651858.ED000142 http://www.ncbi.nlm.nih.gov/pubmed/31643080 https://doi.org/10.1136/bmj.315.7109.629 http://www.ncbi.nlm.nih.gov/pubmed/9310563 https://doi.org/10.1080/08098131.2015.1008558 https://doi.org/10.1080/08098131.2015.1008558 https://doi.org/10.1371/journal.pone.0294529 https://doi.org/10.1371/journal.pone.0294529 http://www.ncbi.nlm.nih.gov/pubmed/37972201 https://humanfactors.jmir.org/2025/1/e70372 Personalized Interactive Music Systems for Physical Activity and Exercise: Exploratory Systematic Review and Meta-Analy- sis JMIR Hum Factors 2025;12:e70372 URL: https://humanfactors.jmir.org/2025/1/e70372 doi: 10.2196/70372 © Andrew Danso, Tiia Kekäläinen, Friederike Koehler, Keegan Knittle, Patti Nijhuis, Iballa Burunat, Pedro Neto, Anasta- sios Mavrolampados, William M Randall, Niels Chr Hansen, Alessandro Ansani, Timo Rantalainen, Vinoo Alluri, Martin Hartmann, Rebecca S Schaefer, Johanna K. Ihalainen, Rebekah Rousi, Kat R Agres, Jennifer MacRitchie, Petri Toiviainen, Suvi Saarikallio, Sebastien Chastin, Geoff Luck. Originally published in JMIR Human Factors (https://humanfactors.jmir.org), 08.09.2025. This is an open-access article distributed under the terms of the Creative Commons Attribution License (https:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Human Factors, is properly cited. The complete bibliographic information, a link to the original publication on https://humanfactors.jmir.org, as well as this copyright and license information must be included. JMIR HUMAN FACTORS Danso et al https://humanfactors.jmir.org/2025/1/e70372 JMIR Hum Factors 2025 | vol. 12 | e70372 | p. 24 (page number not for citation purposes) https://humanfactors.jmir.org/2025/1/e70372 https://doi.org/10.2196/70372 https://humanfactors.jmir.org https://creativecommons.org/licenses/by/4.0/ https://creativecommons.org/licenses/by/4.0/ https://humanfactors.jmir.org https://humanfactors.jmir.org/2025/1/e70372 Personalized Interactive Music Systems for Physical Activity and Exercise: Exploratory Systematic Review and Meta-Analysis Introduction Background Role of Music in Enhancing Physical Activity and Exercise Personalized Interactive Music Systems in Physical Activity and Exercise Methods Study Design Eligibility Criteria Information Sources Search Strategy Selection Process and Data Collection Process Data Extraction Preregistration Deviations Operationalization of Terms Study Risk of Bias Assessment Data Synthesis and Analysis Methods Results Study Selection Study Characteristics Reported Outcome Measures Risk of Bias in Studies PIMSs Used in Experimental Studies PIMSs Used in Proof of Concept and User Testing Studies Meta-Analyses Results for Physical Activity Level Results for Physical Exertion Results for RPE Results for Affective Valence Meta-Regression Analysis Publication Bias Analysis (Egger Test) Discussion Principal Findings Heterogeneity in PIMSs’ Outcomes: Methodological Influences Feasibility of PIMSs on Physical Activity Levels and Affective Outcomes PIMSs’ Tempo Adjustments and Synchronization in Physical Activity and Exercise Outcomes Limitations and Future Directions Conclusions