Sage Journals: Discover world-class research

Abstract

Objective

To evaluate the specific contribution of the real-time postural feedback component during training on balance and mobility in individuals with Parkinson’s disease (PD).

Method

Five databases were searched in June 2025. Studies comparing the same basic balance and gait training with versus without feedback were included. Meta-analysis was conducted and the quality of evidence was assessed.

Result

Eleven studies were included. Narrative analysis indicated that additional feedback may immediately improve dynamic balance and modulate gait speed but not static balance. Meta-Analysis Showed providing feedback significantly augmented post-training improvement in standing balance (SMD = 0.880, 95% CI = 0.345∼1.416, moderate-quality), but not on walking stability (SMD = 0.516, 95% CI = −0.025∼1.058, moderate-quality), Berg Balance Scale (MD = −0.079, 95% CI = −2.159∼2.000, very-low-quality), Timed-Up-and-Go (MD = 0.939, 95% CI = −1.414∼3.293, low-quality), and walking speed (MD = 0.002, 95% CI = −0.057∼0.062, low-quality).

Conclusion

Providing real-time postural feedback may have immediate effects on dynamic balance and walking speed and a post-training effect on standing balance in PD.

Registration

CRD42024522330.

Keywords

real-time postural feedback training balance gait falls Parkinson’s

What This Paper Adds

This is the first systematic review and meta-analysis to specifically summarize the unique contribution of real-time postural feedback during training on balance, mobility, and falls in individuals with Parkinson’s disease.

Adding real-time postural feedback during training may improve the immediate effect (performance during the provision of feedback) of dynamic balance, while all post-training outcomes, except standing balance, did not reach significant changes, potentially due to the reliance on feedback and the mismatch between targeted outcomes and feedback provided.

This study highlights the need for further research with larger sample sizes to strengthen the certainty of the evidence.

Applications of Study Findings

The findings suggest that gradually reducing the real-time postural feedback during the training may help to improve the post-training effect of balance for individuals with Parkinson’s disease.

Future devices designed to provide feedback during training may need to ensure that the feedback is closely aligned with the targeted outcomes.

This study provided a preliminary reference for therapy providers to manage the postural feedback during training to achieve more sustained improvements.

Introduction

Parkinson’s disease (PD) is a rapidly increasing global health issue, currently impacting 6.1 million individuals around the world (Ou et al., 2021). It is a progressive neurodegenerative disorder with the symptoms of resting tremor, muscle rigidity, and postural instability during stance and gait (Allen et al., 2013; Hou et al., 2017).

Due to the difficulties in controlling movements and responding to environmental changes, falls are common in the daily life of the population with PD. A systematic review indicated that 18 to 65% of individuals with PD reported recurrent falls annually, whereas only around 10% of recurrent falls were reported in the general older population each year (Allen et al., 2013; Carpenter et al., 2014). In addition, individuals with PD who had a history of recurrent falls exhibit a notably high annual fall frequency, with reports indicating an average of at least 4.7 falls per year (Allen et al., 2013). This elevated fall frequency undeniably heightens the risk of fall-related injuries among individuals with PD, consequently impacting their quality of life. Furthermore, the direct medical expenses in the United States resulting from falls and related injuries in people with Parkinson’s disease have been estimated at $25.4 billion, with an additional $14.2 million attributed to indirect costs (Appeadu & Gupta, 2023). Therefore, preventing falls in people with Parkinson’s disease is essential for preserving their quality of life and minimizing economic costs.

Falls are commonly associated with postural instability and gait disturbance in individuals with PD (Palmisano et al., 2022). These two disabling motor symptoms have been reported with a low response to medication (Feng et al., 2019; Giardini et al., 2018; Wilson et al., 2020). Recently, physical exercise was shown to improve balance and gait in individuals with PD (Ernst et al., 2023; Feng et al., 2020; Santos et al., 2017). For instance, activities such as Taichi and tango dancing have demonstrated beneficial effects in managing symptoms of Parkinson’s disease. Among various interventions, balance training stands out for its lasting impact, with improvements persisting for as long as 12 months. Additionally, engaging in these exercises has the potential to lower the incidence of falls by as much as 60% (Mak et al., 2017). In addition, treadmill training was also shown to improve the ability of dual-task walking for individuals with PD (Gaßner et al., 2022). These make physical exercise potentially the most effective non-pharmacological approach to improve balance and gait deficits, and prevent falls in individuals with PD (Appeadu & Gupta, 2023; Kim et al., 2013; Smania et al., 2010).

The ability to maintain postural stability and walk effectively depends on the integration of information from multiple sensory systems, including visual cues, somatosensory feedback from muscles and joints, and input from the vestibular system responsible for balance (Giardini et al., 2018). Hence, there is a trend to use sensory augmentation as an assistance for enhancing balance and gait during traditional training. The concept of real-time postural feedback has emerged as a novel approach to augment or replace sensory input in individuals with sensory impairments (Dockx et al., 2016). By delivering instant information on postural movements, this method supports the improvement and coordination of motor skills (Dockx et al., 2016). Advancements in sensor technology and posture detection capabilities have enhanced the ability to monitor body movements, enabling the provision of real-time postural feedback to individuals engaged in balance or mobility exercises (Liang et al., 2024).

Theoretically, with additional real-time postural feedback, individuals with PD may correct posture and attend to their body motions during training and gain further improvement on balance and mobility. Virtual reality training is one of the popular technologies to enhance sensory integration with augmented visual feedback. Lei et al. (2019) published a systematic review that compared the effects between postural feedback-based virtual reality training and traditional training, indicating that virtual reality training had better performance on balance and gait than traditional training in individuals with PD. In addition, a systematic review showed that external sensory-cued therapy could elevate daily living performance after training and a follow-up period of up to 1 year (Cassimatis et al., 2016).

However, the studies included in these reviews tended to compare the feedback-based balance and mobility training versus conventional training or no intervention (Cassimatis et al., 2016; Lei et al., 2019). In addition, research adopting external sensory-cued therapy training tended to offer an instructional cue (e.g., project laser to ground with a given distance from the body and rhythmic cues), rather than cues based on body posture or performance (Harrison et al., 2025; Pallavi et al., 2025). These study designs cannot isolate the effect of real-time postural feedback. The value of incorporating real-time postural feedback within balance or mobility training remains largely unclear (Dockx et al., 2016; Radder et al., 2020; Santos et al., 2019).

Therefore, this systematic review aimed to summarize the latest evidence on the effects of adding real-time postural feedback during balance and mobility training. The postural feedback should focus on the movement performance during the training, such as body sway and walking performance. To delineate the role of feedback, we targeted studies that adopted an experimental group/condition doing a basic balance or gait training with real-time postural feedback, and a control group/comparison condition doing the same basic balance/mobility training but without feedback. The immediate effect of the provision of feedback during a balance/mobility task and the post-training effect after a feedback-instilled training course were evaluated.

Methods

This systematic review was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline (Page et al., 2021; Salamh & and Kolber, 2025). The protocol of this systematic review was pre-registered on the PROSPERO website after the consolidation of the PICO model and search strategy (CRD42024522330).

Sources of Data and Searches

A comprehensive literature search of electronic databases, including PubMed, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Embase, and Scopus, was performed to shortlist relevant articles. Search terms were related to the PICO model (Population, Intervention, Comparison, and Outcomes), of which Population was individuals with PD, Intervention was balance or gait training with concurrent real-time postural feedback, Comparison was a control group adopting the same balance or gait training as the intervention group but without real-time postural feedback, and Outcomes was balance and gait performances during the provision of feedback (immediate effect), or after a course of training (post-training effect), and falls. The full search strategy is shown in the Supplemental Table S1, which was checked and modified by the librarian of the university before the comprehensive search. The initial search was conducted on 11 July 2023, and the last search was performed on 3 June 2025. A forward citation search for all included articles was performed on 2 September 2025 to look for relevant articles.

Titles and abstracts were screened by two researchers (either two from HYC, CMF, SMLA, or SGSL) independently to sort out the relevant studies. Full texts of relevant studies were checked for final inclusion according to the inclusion and exclusion criteria. Any discrepancy during the screening was solved by discussion with the principal supervisor (FMHL).

Study Selection

The inclusion criteria are presented according to the PICOS (Population, Intervention, Comparison, Outcomes, and Study design) framework. Population: people with PD; Intervention: balance or gait training plus immediate postural feedback; Comparison: the same balance and gait training without the feedback component; Outcomes: balance, gait performance or falls; Study Design: (1) within-subject design, comparing the performance of targeted outcomes with and without the provision of feedback while performing the tasks (immediate effect), or (2) randomized controlled trials to illustrate the effect after training with or without feedback provided (post-training effect).

Studies were excluded if they: (1) could not fulfill the inclusion criteria; (2) were protocols; (3) had no full text; (4) were not published in English.

Methodological Quality Assessment and Quality of Evidence

The methodological quality of the included Randomized Controlled Trials (RCTs) was assessed using the PEDro scale (https://pedro.org.au/). The PEDro scores were categorized as follows: excellent (9–10), good (6–8), fair (4–5), and poor (0–3) (Cashin & McAuley, 2020). Additionally, the appraisal tool of the National Heart, Lung, and Blood Institute (NHLBI) study quality assessment was used for the methodological quality of the studies with a within-subject design (Chouksey et al., 2025; NHLBI, 2014). The overall score of this tool was categorized as high (9–12), fair (5–8), and poor quality (0–4) (Hairul Hisham et al., 2025). The methodological quality of all the RCT-designed articles and the studies with a non-RCT design were rated and verified by the two researchers independently (either two of HYC, CMF, SMLA, or SGSL). For the outcomes with a pooled effect size by meta-analysis, a GRADE profile (Grading of Recommendations Assessment, Development and Evaluation) was used to rate the quality of evidence. Specific criteria for rating the quality of evidence are presented in Supplemental Table S2 (Liang et al., 2025).

Data Extraction and Analysis

Original data of each included study were extracted and double-checked by two researchers independently (either two of HYC, CMF, SMLA, or SGSL). Subject characteristics (age, gender, and stage of PD), training dosage (type of feedback, basic training component, duration, and frequency of training), outcome measures, and the main results were retrieved. The stage of Parkinson’s disease (PD) refers to scales that measure the progression of PD in relation to functional disability. The Hoehn and Yahr (H&Y) scale is most commonly used, classifying disease severity into five stages (1–5), with higher stages indicating more severe motor symptoms (Hairul Hisham et al., 2025).

Balance and mobility were the targeted outcome measures in our study. Berg balance scale (BBS) scores, static standing balance (i.e., bipedal and unipedal stance), balance confidence, and fall efficacy scores were used to evaluate overall balance performance. Timed-Up-and-Go (TUG), walking speed, and quality of gait were extracted to examine individuals’ mobility performance.

To evaluate the isolated effect of additional real-time postural feedback on the balance and gait of PD patients, meta-analyses were conducted for outcomes that were measured by at least three articles by using the software of Comprehensive Meta-Analysis (Version 3.3.070). A random-effect model was used for the meta-analysis due to the heterogeneity in study design, subject characteristics, and training settings. Statistical heterogeneity of the meta-analysis was measured by the index of inconsistency (I²). Mean difference (MD) or standardized mean difference (SMD) with their 95% confidence intervals (CI) was computed to illustrate the isolated effect of real-time postural feedback on targeted outcome measures. Results that were not included in the meta-analysis would be synthesized by the narrative analysis to report their study design and findings.

Results

The database search found a total of 2260 records. One hundred and five records remained after the titles and abstracts screening. After the full-text screening, eleven articles (Byl et al., 2015; Carpinella et al., 2017; Ghoseiri et al., 2009; High et al., 2018; Khatavkar et al., 2024; Mayo et al., 2024; McMaster et al., 2022; Nanhoe-Mahabier et al., 2012; Pompeu et al., 2012; Winkler et al., 2022; Yang et al., 2016), fulfilled all the inclusion criteria and were included in this systematic review (Figure 1). No eligible study was found in the citation searching. A total of 226 individuals with PD were included across these articles.

Figure 1.

PRISMA flow diagram for studies selection

Characteristics of the Selected Studies

The mean ages of individuals ranged from 58.6 to 75.6 years. Disease severity ranged from Hoehn and Yahr (H&Y) stages 1 to 4, encompassing participants from minimal unilateral involvement (stage 1) to severe disability with preserved ability to stand or walk unassisted (stage 4). Four studies (Ghoseiri et al., 2009; High et al., 2018; Khatavkar et al., 2024; McMaster et al., 2022) adopted a within-subject design that investigated the immediate effect by comparing the balance and mobility performance across the conditions of with and without real-time postural feedback, whereas seven studies (Byl et al., 2015; Carpinella et al., 2017; Mayo et al., 2024; Nanhoe-Mahabier et al., 2012; Pompeu et al., 2012; Winkler et al., 2022; Yang et al., 2016) were randomized controlled trials evaluating the post-training effect.

Among the studies investigating the immediate effect, three studies underwent the walking or balance task with vibrotactile postural feedback on heel time, stride angle of ankle joint, and body sway that was provided by wearable devices, including a vibratory waist belt (Ghoseiri et al., 2009; High et al., 2018) or an instrumented shoe and vibratory insole (Khatavkar et al., 2024), one study provided visual feedback of mediolateral lean of the trunk during walking by a screen located at the end of the walkway (McMaster et al., 2022).

For the studies that evaluated post-training effect, three studies provided pure visual postural feedback via the screen of the training system, including a balance system with a force plate (Yang et al., 2016), Nintendo Wii (Pompeu et al., 2012), and a self-developed system with smart shoes and pants for postural monitoring (Byl et al., 2015). One study (Nanhoe-Mahabier et al., 2012) delivered pure vibratory feedback on trunk sway via a headband, and one study (Mayo et al., 2024) offered pure acoustic feedback for a good step, defined as having sufficient angular velocity at the ankle joint, that was detected by smart shoes with sensors. Multicomponent postural feedback incorporated acoustic with visual feedback (Carpinella et al., 2017), or vibrotactile feedback was adopted by two studies (Winkler et al., 2022).

Details of participants and training protocols of the included studies are provided in Table 1. The summary of the characteristics of the included studies and the results details are presented in the Supplemental Table S3 & Table S4.

Table 1.

Study Characteristics of Shortlisted Studies

Study label	Subject characteristics				Intervention group/experimental condition					Comparison group/control condition
Study label	Sample size	Mean age ± SD	Gender	H&Y scale (PD medication: On/off)	Type of feedback	Basic training	Duration of program	Duration per session	Frequency of sessions	Comparison group/control condition
Ghoseiri et al. (2009)	Total: 14 within-subject design	59.9 ± 9.1	12 males	2–4 (on)	• Contents of feedback: Vibrotactile feedback while the lower limb is stepping on the ground, which is detected by foot micro-switches	Normal walking	1 day	NA, single-group assessment compared conditions with/without feedback		Normal walking without feedback
Ghoseiri et al. (2009)	Total: 14 within-subject design	59.9 ± 9.1	2 females	2–4 (on)	• Way of delivering feedback: Vibratory waist belt	Normal walking	1 day			Normal walking without feedback
High et al. (2018)	Total: 9 within-subject design	69 ± 10.25	7 males	UPDRS III: 25.22 ± 13.24 (on)	• Contents of feedback: Vibrotactile feedback for the body sway	Balance training: feet together in eyes open/closed on firm and foam surface; tandem stance with eyes open on firm surface	1 day	NA, single-group assessment compared conditions with/without feedback		Same as the experimental condition, but without feedback
High et al. (2018)	Total: 9 within-subject design	69 ± 10.25	2 females	UPDRS III: 25.22 ± 13.24 (on)	• Way of delivering feedback: Vibrotactile feedback waist belt		1 day			Same as the experimental condition, but without feedback
McMaster et al. (2022)	Total: 24 within-subject design	68 ± 7.6	18 males	1–3 (on)	• Contents of feedback: Visual feedback for the mediolateral lean of the trunk detected by VICON system during walking	Normal walking	1 day	NA, single-group assessment compared conditions with/without feedback		Walking without feedback
McMaster et al. (2022)	Total: 24 within-subject design	68 ± 7.6	6 females	1–3 (on)	• Way of delivering feedback: Screen in front of the subjects	Normal walking	1 day			Walking without feedback
Khatavkar et al. (2024)	Total: 8 Within-subject design	66.12 ± 7.03	8 males	UPDRS III: 43.25 ± 12.47 (on)	• Contents of feedback: Vibrotactile (haptic) feedback for the foot strike angle of ankle joint above a threshold during walking	Walking on a 6-m walkway back and forth for a 3-min duration	1 day	NA, single-group assessment compared conditions with/without feedback		Walking training without the feedback while wearing the same instrumented shoe and insole
Khatavkar et al. (2024)	Total: 8 Within-subject design	66.12 ± 7.03	8 males	UPDRS III: 43.25 ± 12.47 (on)	• Way of delivering feedback: Instrumented shoe and insole		1 day
Pompeu et al. (2012)	Total: 32	67.4 ± 8.1	17 males	1–2 (on)	• Contents of feedback: Visual feedback involves the body and limb location during the training	30 min of global exercises include warming-up, strengthening exercises, and diagonal patterns exercise for the trunk, neck, and limbs; 30 min of balance exercises (static balance, dynamic balance, and stationary gait)	7 weeks	1 h	2 times/week	Same global exercise and balance exercise without feedback
	IG: 16		15 females		• Way of delivering feedback: Nintendo Wii games with visual feedback on screen
	CG: 16		15 females
Nanhoe-Mahabier et al. (2012)	Total: 20	IG:	IG:	Mean ± SD 1.6 ± 0.1 (on)	• Contents of feedback: Vibrotactile feedback for trunk sway	6 tasks: standing with feet together with eyes closed, standing on one leg with eyes open, tandem standing with eyes closed, standing with feet together on foam with eyes closed, walking 9 m at preferred speed with eyes open and walking 15 tandem steps with eyes closed	1 day	No reported	Not reported	Same basic training without feedback
	IG:10	59.3 ± 2.0	8 males		• Way of delivering feedback: BalanceFeedom system, which consists of a waist belt (lower trunk) with angular velocity sensors, and a headband providing feedback
	CG:10	CG:	2 females
		58.6 ± 2.5	CG:
			8 males
			2 females
Byl et al. (2015)	Total: 12	IG:	IG:	1–3 (NR)	• Contents of feedback: Visual kinematic feedback involves the lower limb’s location, gait events, joint angles, and force on the ground during training, which is detected by the smart shoes and pants	Gait training involves various specific activities, including sit-to-stand, balance exercises, walking ± dual task, stand and swing phase exercises, coordination training, stairs walking, plyometric exercises, strengthening, and flexibility exercises. Such training is done both indoors and outdoors, on the ground and on the treadmill	6 weeks	30 min	2 times/week	Same basic training while wearing usual shoes and pants.
	IG: 7	68.5 ± 3.6	3 males		• Way of delivering feedback: The computer screen
	CG: 5	CG:	4 females
		70 ± 2.9	CG:
			4 males
			1 female
Yang et al. (2016)	Total: 20	72.5 ± 8.4	14 males	2–3 (on)	• Contents of feedback: Visual feedback involves the center of gravity of the body by a force plate	10 min of static posture exercise and 20 min of dynamic balance training	6 weeks	50 min	2 times/week	Same as the movement of the intervention group.
	IG: 11 to 10		9 females		• Way of delivering feedback: Balance system with a screen
	CG: 12 to 10		9 females		• Way of delivering feedback: Balance system with a screen
Carpinella et al. (2017)	Total: 42	IG:	CG:	2–4 (on)	• Contents of feedback: Visual and auditory feedback involves the body and lower limbs’ location/angle and task performance during training	Walking, stance balance, and stand-up exercises	7 weeks	45 min	3 times/week	Basic training that was similar to that of IG but without feedback.
	IG: 22	73.0 ± 7.1	14 males		• Way of delivering feedback: Gamepad system with screen and sound feedback
	CG: 20	CG:	3 females
		75.6 ± 8.2	IG:
			9 males
			11 females
Winkler et al. (2022)	Total: 18	IG:	IG:	1–4 (on)	• Contents of feedback: Multimodal sensory feedback involves somatosensory and acoustic feedback for stepping on ground motion, visual feedback for trunk verticality, and proprioceptive feedback for arms’ sense of location/angulation	Gait, stepping forward and backward quickly, and full stride walking exercises. Balance training such as single leg stance, turning 90° in stepping, is done with or without the addition of foam and with eyes open or closed	18 weeks (6-week intervention, 6-week without ex, 6-week for examining retention effects)	Up to 45 min, but were reduced with practice to 20–30 min	5 times/week	Same basic exercise
	IG: 15 to 11	67.4 ± 6.7	6 males		• Way of delivering feedback: Balance Matters® System, that equipped with movable foot plates and clickers at the heel and forefoot areas, provides somatosensory and acoustic feedback for foot during stepping motion. Visual feedback: a laser light that is attached to the individual’s chest. Proprioceptive feedback: 1-pound weight on the wrist and forefeet to promote arm swing and hip flexion
	CG: 11 to 7	CG:	5 females
		68.6 ± 5.6	CG:
			4 males
			3 females
Mayo et al. (2024)	Total: 27	IG:	IG:	2–3 (on)	• Contents of feedback: Auditory feedback while the steps during walking is detected as “good step,” by a sensor on shoe measuring the angular velocity	10 to 15 repetitions of exercise movement for a better walking pattern.	3 months	At least 5 min twice a day	7 days/week	Participants walked while wearing the sensor shoe without feedback
	IG: 18 to 14	70.2 ± 8.5	4 females		• Way of delivering feedback: The sensor shoe provided feedback
	CG: 9 to 7	CG:	13 males
		70.7 ± 8.8	CG:
			3 females
			6 males

Notes: H&Y = Hoehn & Yahr Scale. SD = Standard Deviation. IG = Intervention Group. CG = Comparison Group. NA = Not Applicable. NR = Not Reported. UPDRS: Unified Parkinson’s Disease Rating Motor Scale.

The Methodological Quality of Studies and Evidence of Quality

The methodological quality of the included studies with a within-subject design (Ghoseiri et al., 2009; High et al., 2018; Khatavkar et al., 2024; McMaster et al., 2022) was rated as high quality (low risk of bias) with a score of 9 to 10 (Supplemental Table S5). No justification for the sample size calculation (Ghoseiri et al., 2009; High et al., 2018; Khatavkar et al., 2024), no blinding of assessors (Ghoseiri et al., 2009; High et al., 2018; Khatavkar et al., 2024; McMaster et al., 2022), and the lack of multiple measurements (Ghoseiri et al., 2009; High et al., 2018; Khatavkar et al., 2024; McMaster et al., 2022) were the main concerns of downgrading the methodological quality of the study (Supplemental Table S5). For the randomized control trials, five (Carpinella et al., 2017; Mayo et al., 2024; Nanhoe-Mahabier et al., 2012; Winkler et al., 2022; Yang et al., 2016) out of seven (Byl et al., 2015; Carpinella et al., 2017; Mayo et al., 2024; Nanhoe-Mahabier et al., 2012; Pompeu et al., 2012; Winkler et al., 2022; Yang et al., 2016) were rated as good methodological quality (PEDro score ≥ 6). Two studies (Byl et al., 2015; Pompeu et al., 2012) had fair methodological quality (PEDro score 4–5) (Table 2). Most of the included RCTs did not fulfill the criteria of concealed allocation, blinding of subjects, and therapists (Byl et al., 2015; Carpinella et al., 2017; Mayo et al., 2024; Nanhoe-Mahabier et al., 2012; Pompeu et al., 2012; Winkler et al., 2022; Yang et al., 2016).

Table 2.

The Methodological Quality of Randomized Controlled Trials by the Scale of Physiotherapy Evidence Database

Criteria/Study label	Byl et al. (2015)	Winker et al. (2022)	PomPeu et al. (2012)	Nanhoe-Mahabier et al. (2012)	Carpinella et al. (2017)	Yang et al. (2016)	Mayo et al. (2024)
Eligibility criteria	Y	Y	N	Y	Y	Y	Y
Random allocation	Y	Y	Y	Y	Y	Y	Y
Concealed allocation	N	N	N	N	N	N	Y
Baseline comparability	N	Y	Y	Y	Y	Y	Y
Blind subjects	N	N	N	N	N	N	N
Blind therapists	N	N	N	N	N	N	N
Blind assessors	N	Y	Y	N	Y	Y	Y
Adequate follow-up	Y	N	N	Y	Y	Y	N
ITT analysis	N	Y	N	Y	N	Y	Y
Between-group effect	Y	Y	Y	Y	Y	Y	N
Point estimates	Y	Y	Y	Y	Y	Y	Y
Total score	4	6	5	6	6	7	6

Notes: ITT, Intention-to-treat. Y, the study fulfilled the criterion. N, the study did not fulfill the criterion.

Byl et al. (2015).

Winkler et al. (2022).

Pompeu et al. (2012).

Nanhoe-Mahabier et al. (2012).

Carpinella et al. (2017).

Yang et al. (2016).

Mayo et al. (2024).

For the immediate effect of providing feedback, none of the outcomes of interest was measured by at least three studies. Therefore, no meta-analysis could be conducted. For the post-training effect of providing feedback, at least three studies have measured outcomes on balance and mobility. Meta-analyses and rating on the quality of evidence were thus conducted. The quality of evidence of all eligible outcomes was graded down by imprecision due to the small sample size of the study and the 95% confidence intervals of the pooled effect size spanning plausible benefit and/or harm. The evidence generated by this review for standing balance and gait stability was of moderate quality, while the evidence for the timed-up-and-go test and the gait speed was of low quality. The evidence for the Berg Balance scale was of very low quality (Table 3).

Table 3.

Quality of Evidence for Outcomes by Grade of Recommendation, Assessment, Development, and Evaluation Approach

Outcome	Risk of bias	Inconsistency	Imprecision	LargeEffect	Pooled effect size (standardized as SMD for GRADE quality rating)	GRADE quality
Berg balance Scale	0	−1^a	−2^b	0	SMD: −0.004 [−0.648, 0.641]	Very low
Standing balance	−1^c	0	−1^d	+1^e	SMD: 0.880 [0.345, 1.416]	Moderate
Time-up-and-go	0	0	−2^b	0	SMD: 0.186 [−0.280, 0.653]	Low
Gait Speed	0	0	−2^b	0	SMD: 0.037 [−0.457, 0.530]	Low
Gait Stability	0	0	−1^d	0	SMD: 0.516 [−0.025, 1.058]	Moderate

Notes: SMD, standardized mean difference/Cohen’s d.

^aHeterogeneity I² larger than 50% in the meta-analysis.

^bThe 95% CI spanned −0.2 and 0.2.

^cMore than half of the participants included in the primary analysis from trials with a PEDro score less than 6.

^dThe number of participants included in the primary meta-analysis was less than 400 or the 95% CI spanned 0/0.2.

^ePooled effect size larger than 0.8.

Immediate Effect of Real-Time Postural Feedback on Balance and Gait

Four studies (Ghoseiri et al., 2009; High et al., 2018; Khatavkar et al., 2024; McMaster et al., 2022) (n = 45) investigated the immediate effect of adding real-time postural feedback during static and dynamic balance (High et al., 2018; Khatavkar et al., 2024; McMaster et al., 2022) or gait speed (Ghoseiri et al., 2009; McMaster et al., 2022).

Only one study, which was rated as high quality, assessed static balance and reported an insignificant result. High et al. (2018) (n = 9) measured the static standing balance with the postural sway of the center of pressure (i.e., path length, sway velocity, sway area) under the condition with vibrotactile feedback or without feedback. The vibrotactile feedback was provided to the individuals with PD, when the sway from the center of their base of support exceeded 10% during balance tasks. No significant difference was found across conditions.

In contrast, two studies that were rated as high quality reported significant improvement in dynamic balance. McMaster et al. (2022) measured the mediolateral trunk lean (degree) during walking, showing that providing real-time visual feedback of the mediolateral lean of the trunk during walking to individuals with PD significantly reduced the mediolateral lean of the trunk during walking, when compared with the walking condition without feedback (3.48 versus 4.19°, p < 0.05, n = 24). In addition, Khatavkar et al. (2024) reported that walking on a 6-m walkway back and forth for 3 min with the vibrotactile (haptic) feedback on foot given when the foot strike angle above a threshold, significantly improved the foot strike angle (2.93 versus −5.43°, p = 0.0362, n = 8) and toe clearance distance (94.29 versus 87.60, p = 0.0019, n = 8) during walking, compared with the condition of the same walking without walking performance feedback.

For mobility outcome, two studies that were rated as high quality measured the walking time of a 10-m pathway or the walking velocity, and illustrated the significant but opposite results. Ghoseiri et al. (2009) indicated that the additional real-time heel-contact events sensory feedback by vibration on the waist during walking significantly improved gait speed compared with the walking condition without feedback (10.72 versus 12.14 s, p = 0.001, n = 14). However, McMaster et al. (2022) showed that the gait speed significantly decreased under the condition of visual feedback of body alignment (mediolateral lean of the trunk) when compared with the condition without visual feedback during walking (1.01 versus 1.12 m/s, p < 0.05, n = 24).

Post-training Effect of Real-Time Postural Feedback on Balance

Balance performance was measured in five randomized controlled trials mainly by the Berg Balance Scale (BBS) (Byl et al., 2015; Carpinella et al., 2017; Pompeu et al., 2012; Yang et al., 2016), and standing body sway or velocity (i.e., unipedal stance test and double leg stance test) (Carpinella et al., 2017; Nanhoe-Mahabier et al., 2012; Pompeu et al., 2012).

For the Berg Balance Scale, four studies were included in the meta-analysis. It revealed that adding real-time postural feedback during balance training (Carpinella et al., 2017; Pompeu et al., 2012; Yang et al., 2016) or gait training (Byl et al., 2015; Carpinella et al., 2017) had no difference relative to the training without feedback in individuals with PD on the Berg Balance Scale (MD = −0.079, 95% CI = −2.159 to 2.000, I² = 63.85, n = 104, Figure 2a), with very low quality of evidence.

Figure 2.

Results of the meta-analyses; (a) Berg Balance Scale (n = 104); (b) Standing balance (n = 89); (c) Time-Up-and-Go Test (n = 72); (d) Gait speed (n = 67); (e) Gait stability (n = 55)

For standing balance (Carpinella et al., 2017; Nanhoe-Mahabier et al., 2012; Pompeu et al., 2012), the meta-analysis with three included RCTs indicated that the real-time postural feedback group had a significant positive effect on standing balance when compared with non-feedback training (SMD = 0.880, 95% CI = 0.345 to 1.416, I² = 30.81, n = 89, Figure 2b) in individuals with PD, with moderate quality of evidence.

Post-Training Effect of Real-Time Postural Feedback on Mobility

Mobility performance was measured by six studies (Byl et al., 2015; Carpinella et al., 2017; Mayo et al., 2024; Nanhoe-Mahabier et al., 2012; Winkler et al., 2022; Yang et al., 2016) using the Timed-Up-and-Go test (TUG), gait speed, gait stability, and standardized walking obstacle course test.

For TUG, this functional mobility test was measured by three studies (Byl et al., 2015; Carpinella et al., 2017; Yang et al., 2016), which were included in the meta-analysis. It showed that the additional real-time postural feedback has no significant effect on TUG performance (Byl et al., 2015; Carpinella et al., 2017; Yang et al., 2016) (MD = 0.939, 95% CI = −1.414 to 3.293, I² = 0, n = 72, Figure 2c), with low quality of evidence. Besides, Winkler et al. (2022) reported an insignificant result on dual-task TUG.

Gait speed was measured by three studies (Byl et al., 2015; Carpinella et al., 2017; Winkler et al., 2022). The meta-analysis showed that no effect was observed after adding real-time postural feedback during training on gait speed when compared with training without feedback (MD = 0.002, 95% CI = −0.057 to 0.062, I² = 5.205, n = 67, Figure 2d), with low quality of evidence.

Gait stability was measured in three studies (Byl et al., 2015; Nanhoe-Mahabier et al., 2012; Yang et al., 2016) by the Dynamic Gait Index (DGI) (Byl et al., 2015; Yang et al., 2016) and the roll angle during walking (Nanhoe-Mahabier et al., 2012). The meta-analysis showed an insignificant result after adding real-time postural feedback when compared with training without feedback (SMD = 0.516, 95% CI = −0.025 to 1.058, I² = 0, n = 55, Figure 2e), with moderate quality of evidence.

Other mobility outcomes reported included walking obstacle course and 6-min walking test with contrasting results (Mayo et al., 2024). A significantly worse obstacle course performance was noted after the walking training with feedback, compared with the walking training without feedback (with feedback: 4.3 seconds; without feedback: −6.1 s, n = 22). In contrast, the group that received walking training with feedback showed a significantly greater improvement in six-minute walk test performance (with feedback: 66.4 m; without feedback: −19.4 m, n = 20).

Post-training Effect of Real-Time Postural Feedback on Falls

Falls-related outcome was reported by only one RCT using the Activities-Specific Balance Confidence Scale (ABC) (Carpinella et al., 2017). The study adopted virtual and auditory real-time postural feedback during the balance and walking training, indicating an insignificant effect on ABC (p = 0.186) after training.

Discussion

This systematic review was the first review that summarized evidence on the isolated effect of adding real-time postural feedback during balance or mobility training on balance, mobility, and falls in individuals with PD. The results showed that real-time postural feedback boosted the immediate performance of dynamic balance and modulated walking speed with no effect on static balance. For post-training effect, adding real-time postural feedback only significantly improved standing balance with no effect on mini-BESTest, TUG, gait speed, and gait stability. Furthermore, with only a single study reporting a non-significant effect of ABC, the effect of real-time postural feedback on falls-related outcomes remains inconclusive. With the heterogeneity observed in the results, the discussion of this review will be centered around the potential reasons for the heterogeneity of the results.

Outcome-Matching Feedback Could be Important for a Significant Immediate Effect

Our results suggested that additional real-time postural feedback may improve performance on dynamic balance immediately. This aligned with mechanistic studies showing that the external feedback or cues of the performance enhanced the functional connectivity between the supplementary motor area (Zadeh et al., 2024), which is the foundational structure of the complex motor planning and motor coordination (Nachev et al., 2008).

However, this enhancement was not observed in static balance and gait speed. A potential reason for this phenomenon could be the mismatched feedback content with the targeted outcomes. In general, outcomes that do not align with the feedback provided have reported unfavorable immediate effects (McMaster et al., 2022). In contrast, outcomes that demonstrated favorable immediate effects were often coupled with outcome-matching, specific feedback (Ghoseiri et al., 2009; Khatavkar et al., 2024; McMaster et al., 2022). For instance, providing feedback on trunk lean angle during walking could improve dynamic balance during walking but not gait speed (McMaster et al., 2022). The only exception to this observation came from a study that failed to improve standing balance even by providing outcome-matching feedback, which was potentially due to a small sample size (n = 9) (High et al., 2018). Therefore, it appears that outcome-matching feedback could be important for the improvement observed.

To date, there is no review article that focuses on the real-time postural feedback on immediate effects in individuals with PD directly. Nevertheless, our findings generally align with other similar reviews recently published. For example, a recent narrative review pooled results from four studies that combined immediate and post-training effects of real-time electromyographic biofeedback in Parkinson’s disease (PD) and reported overall benefits on motor outcomes, including postural instability (Diotaiuti et al., 2025). Another systematic review has demonstrated that the immediate effect of dynamic balance was improved by external real-time postural feedback in healthy older adults (Liang et al., 2024).

Overall, considering only four studies included for immediate effect with a small sample size, even though the studies were rated as high quality, a strong conclusion cannot be drawn. Real-time postural feedback appears to be able to improve balance and gait performance when the feedback provided matches with the targeted outcomes.

Gradual Reduction in the Provision of Feedback May Lead to Better Post-Training Outcome

The meta-analyses suggested that additional feedback improved standing balance, but not BBS, the comprehensive balance performance evaluation scale, and other mobility outcomes including TUG, gait speed, and gait stability.

The heterogeneity of the post-training effect could be potentially explained by the different dosages of the real-time postural feedback component during balance or walking training. The only two (Carpinella et al., 2017; Nanhoe-Mahabier et al., 2012) studies that have reported significant improvement both reduced the provision of real-time postural feedback over the training course. Nevertheless, all insignificant studies have provided feedback continuously without modulation during training (Byl et al., 2015; Mayo et al., 2024; Pompeu et al., 2012; Winkler et al., 2022; Yang et al., 2016). This may explain why the improvement was observed only in the meta-analysis of standing balance: it is the only outcome assessed in both studies that reduced feedback provision during training (Carpinella et al., 2017; Nanhoe-Mahabier et al., 2012).

Indeed, previous studies have suggested that providing frequent external feedback during task performance could potentially disrupt learning through the implicit mechanism in the long term (Maxwell et al., 2001; Orrell et al., 2006a, 2006b), which is important for improving balance performance (Orrell et al., 2006a). With external feedback provided throughout the training course, participants could have accommodated to completing balance or mobility tasks under the provision of postural feedback, and gradually developed reliance. However, after the training sessions, the improvement based on the feedback provision would be eliminated in the post assessment, as no feedback was provided during the assessment tasks.

Therefore, it appears that the initial feedback provided insight into the improvement required. Gradually reducing real-time feedback hereafter may allow individuals with PD to learn motor skills and wean off the reliance on extrinsic feedback (Liang et al., 2024), potentially leading to a significant post-training improvement. Given that the level of evidence was moderate for the meta-analysis of static balance, it suggested that additional real-time postural feedback may improve the post-training effect of static balance. However, this needs to be confirmed by larger studies.

Outcome-Matching Feedback Could Also be Essential for Post-Training Improvement

Another potential explanation for the heterogeneity of the meta-analyses could be the mismatch between the feedback provided and the targeted outcome. Postural feedback, such as body sway and limb location, was found to significantly improve standing balance, which is highly relevant to the feedback provided (Carpinella et al., 2017; Nanhoe-Mahabier et al., 2012; Pompeu et al., 2012). However, postural feedback is consistently unable to improve mobility reflected in speed-related measures, including TUG and gait speed, which do not match with the nature of the feedback provided (Byl et al., 2015; Carpinella et al., 2017; Mayo et al., 2024; Winkler et al., 2022; Yang et al., 2016).

A recent systematic review investigated the effect of haptic cues on gait in individuals with PD and indicated the improvement in gait speed (Maddumage Dona et al., 2025), which differs from our review finding. It may be attributed to that the rhythmic cue emphasizes the improvement of automaticity, whereas postural feedback in our review focuses on the segment alignment during the task. Therefore, rhythmic cue could improve walking speed, while postural feedback, which was included in the present review, improves postural stability but not speed.

Overall, our review suggested that the real-time postural feedback provided may need to match the outcomes targeted for improvement of the post-training effect. Given the limited number of included studies, further studies with a large sample size are needed to verify.

Practical Implications

Our review shows that providing real-time postural feedback might improve the immediate effect of dynamic balance in people with various severities of Parkinson’s disease (H&Y stage 1–3/UPDRS III 22-61) if the feedback provided matches the intended outcomes. For the feedback provision, the visual and vibratory feedback were shown to be effective, delivered by dedicated training systems or wearable devices.

Incorporating feedback in balance or walking training may also augment the improvement post-training in people with various severities of Parkinson’s disease (H&Y stage 1–4). The feedback provided may need to match closely with the target outcome. It appears that the intensity of the feedback, provided by a vibratory, auditory or combination of visual and auditory stimuli, may need to be gradually reduced to achieve a post-training improvement.

Given the limited included studies and the limitations of the study design, a strong practical recommendation cannot yet be made, and these preliminary findings warrant further investigation and validation.

Limitations

The systematic review has several limitations. Firstly, most of the real-time postural feedback of the included studies was provided to participants with PD by vision or a multi-component sensory model. The results of this review cannot be generalized to other feedback delivery models. Secondly, despite our search keywords containing falls, with only one study reporting the falls-related outcomes (ABC), we cannot conclude the post-training isolated effect of real-time postural feedback during balance or gait training on falls. Thirdly, the small sample size of the included studies (range: 8 to 42) may lead to insufficient power to detect significant changes in immediate effect (e.g., static standing balance) and post-training effect (e.g., walking stability). Finally, the quality of evidence for outcomes was rated as very low to moderate. The results may change as the new study with a large sample size is included in the future.

Conclusion

Our findings suggest that providing real-time postural feedback may have an immediate effect on improving dynamic balance and modulating gait speed. Adding postural feedback to a course of balance training appears to augment the improvement in standing balance. The provision of feedback may need to match with the targeted outcomes to yield significant immediate and post-training improvement. Gradually reducing the postural feedback provided over the course of training may help to improve its post-training effect. Given the limited studies investigating the isolated effect of real-time postural feedback and the very low to moderate quality of the evidence for the meta-analysis, further research with a large sample size is warranted.

Supplemental Material

Supplemental Material—Does Adding Real-Time Postural Feedback During Training Boost Balance and Mobility in Individuals With Parkinson’s Disease? A Systematic Review and Meta-Analysis

Supplemental Material for Does Adding Real-Time Postural Feedback During Training Boost Balance and Mobility in Individuals With Parkinson’s Disease? A Systematic Review and Meta-Analysis by Guoshi Liang, Hou Yee Ching, Chiu Man Fu, Sin Man Agnes Lau, Wayne Lap Sun Chan, Meizhen Huang, Freddy Man-Hin Lam in Journal of Applied Gerontology

Footnotes

ORCID iDs

Guoshi Liang

Wayne Lap Sun Chan

Meizhen Huang

Freddy Man-Hin Lam

Ethical Considerations

Ethics approval and informed consent were not needed as the research resources were the literature.

Author Contributions

Guoshi Liang: Data curation, Formal analysis, Investigation, Writing—original draft.

Hou Yee Ching: Data curation, Formal analysis, Writing—original draft.

Chiu Man Fu: Data curation, Formal analysis, Writing—original draft.

Sin Man Agnes Lau: Data curation, Formal analysis, Writing—original draft.

Wayne Lap Sun Chan: Conceptualization, Validation, Writing—review & editing.

Meizhen Huang: Conceptualization, Methodology, Writing—review & editing.

Freddy Man-Hin Lam: Conceptualization, Methodology, Supervision, Writing—review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Start-up Fund of The Hong Kong Polytechnic University (P0034487, 2020); and the Departmental General Research Fund of The Hong Kong Polytechnic University (P0042679, 2022).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

Data are available from the corresponding author on reasonable request.*

Footnote

This study followed the guidance of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and was registered in PROSPERO (CRD42024522330).

Supplemental Material

Supplemental material for this article is available online.

References

Allen

N. E.

Schwarzel

A. K.

Canning

C. G.

(2013). Recurrent falls in Parkinson’s disease: A systematic review. Parkinson's Disease, 2013(2), 906274. https://doi.org/10.1155/2013/906274

Appeadu

M. K.

Gupta

(2023). Postural instability. StatPearls Publishing.

Byl

Zhang

Coo

Tomizuka

(2015). Clinical impact of gait training enhanced with visual kinematic biofeedback: Patients with Parkinson’s disease and patients stable post stroke. Neuropsychologia, 79(Pt B), 332–343. https://doi.org/10.1016/j.neuropsychologia.2015.04.020

Carpenter

C. R.

Avidan

M. S.

Wildes

Stark

Fowler

S. A.

A. X.

(2014). Predicting geriatric falls following an episode of emergency department care: A systematic review. Academic Emergency Medicine, 21(10), 1069–1082. https://doi.org/10.1111/acem.12488

Carpinella

Cattaneo

Bonora

Bowman

Martina

Montesano

Ferrarin

(2017). Wearable sensor-based biofeedback training for balance and gait in Parkinson disease: A pilot randomized controlled trial. Archives of Physical Medicine and Rehabilitation, 98(4), 622–630. https://doi.org/10.1016/j.apmr.2016.11.003

Cashin

A. G.

McAuley

J. H.

(2020). Clinimetrics: Physiotherapy evidence database (PEDro) scale. Journal of Physiotherapy, 66(1), 59. https://doi.org/10.1016/j.jphys.2019.08.005

Cassimatis

Liu

K. P. Y.

Fahey

Bissett

(2016). The effectiveness of external sensory cues in improving functional performance in individuals with Parkinson’s disease: A systematic review with meta-analysis. International Journal of Rehabilitation Research, 39(3), 211–218. https://doi.org/10.1097/mrr.0000000000000171

Chouksey

Choure

Singhal

(2025). Health-related quality of life after prosthodontic rehabilitation in patients with COVID-19-associated mucormycosis: A systematic review. Journal of Indian Prosthodontic Society, 25(1), 30–39. https://doi.org/10.4103/jips.jips_202_24

Diotaiuti

Marotta

Vitiello

Di Siena

Palombo

Langiano

Ferrara

Mancone

(2025). Biofeedback for motor and cognitive rehabilitation in Parkinson's disease: A comprehensive review of non-invasive interventions. Brain Sciences, 15(7), 720. https://doi.org/10.3390/brainsci15070720

10.

Dockx

Bekkers

E. M. J.

Van den Bergh

Ginis

Rochester

Hausdorff

J. M.

Mirelman

Nieuwboer

(2016). Virtual reality for rehabilitation in Parkinson’s disease. Cochrane Database of Systematic Reviews, 12(12), CD010760. https://doi.org/10.1002/14651858.CD010760.pub2

11.

Ernst

Folkerts

A. K.

Gollan

Lieker

Caro-Valenzuela

Adams

Cryns

Monsef

Dresen

Roheger

Eggers

Skoetz

Kalbe

(2023). Physical exercise for people with Parkinson’s disease: A systematic review and network meta-analysis. Cochrane Database of Systematic Reviews, 1(1), CD013856. https://doi.org/10.1002/14651858.CD013856.pub2

12.

Feng

Liu

Wang

Gan

Shang

(2019). Virtual reality rehabilitation versus conventional physical therapy for improving balance and gait in Parkinson’s disease patients: A randomized controlled trial. Medical Science Monitor, 25(1), 4186–4192. https://doi.org/10.12659/msm.916455

13.

Feng

Y. S.

Yang

S. D.

Tan

Z. X.

Wang

M. M.

Xing

Dong

Zhang

(2020). The benefits and mechanisms of exercise training for Parkinson’s disease. Life Sciences, 245(1), 117345. https://doi.org/10.1016/j.lfs.2020.117345

14.

Gaßner

Trutt

Seifferth

Friedrich

Zucker

Salhani

Adler

Winkler

Jost

W. H.

(2022). Treadmill training and physiotherapy similarly improve dual task gait performance: A randomized-controlled trial in Parkinson’s disease. Journal of Neural Transmission, 129(9), 1189–1200. https://doi.org/10.1007/s00702-022-02514-4

15.

Ghoseiri

Forogh

Sanjari

M. A.

Bavi

(2009). The effect of a vibratory lumber orthosis on walking velocity in patients with Parkinson’s disease. Prosthetics and Orthotics International, 33(1), 82–88. https://doi.org/10.1080/03093640802647094

16.

Giardini

Nardone

Godi

Guglielmetti

Arcolin

Pisano

Schieppati

(2018). Instrumental or physical-exercise rehabilitation of balance improves both balance and gait in Parkinson’s disease. Neural Plasticity, 2018(10), 5614242. https://doi.org/10.1155/2018/5614242

17.

Hairul Hisham

H. I.

Lim

S. M.

Neoh

C. F.

Abdul Majeed

A. B.

Shahar

Ramasamy

(2025). Effects of non-pharmacological interventions on gut microbiota and intestinal permeability in older adults: A systematic review: Non-pharmacological interventions on gut microbiota/barrier. Archives of Gerontology and Geriatrics, 128, 105640. https://doi.org/10.1016/j.archger.2024.105640

18.

Harrison

E. C.

Tueth

L. E.

Haussler

A. M.

Rawson

K. S.

Earhart

G. M.

(2025). Personalized auditory rhythmic cues to optimize gait in older adults and people with Parkinson disease. Journal of Neurologic Physical Therapy, 49(3), 162–170. https://doi.org/10.1097/npt.0000000000000508

19.

High

C. M.

McHugh

H. F.

Mills

S. C.

Amano

Freund

J. E.

Vallabhajosula

(2018). Vibrotactile feedback alters dynamics of static postural control in persons with Parkinson’s disease but not older adults at high fall risk. Gait and Posture, 63(1), 202–207. https://doi.org/10.1016/j.gaitpost.2018.05.010

20.

Hou

Chen

Liu

Qiao

Zhou

F.-M.

(2017). Exercise-induced neuroprotection of the nigrostriatal dopamine system in Parkinson’s disease. Frontiers in Aging Neuroscience, 9(1), 358. https://doi.org/10.3389/fnagi.2017.00358

21.

Khatavkar

Tiwari

Bhat

Joshi

(2024). Investigating the effects of a kinematic gait parameter-based haptic cue on toe clearance in Parkinson’s patients. Annals of Biomedical Engineering, 52(8), 2039–2050. https://doi.org/10.1007/s10439-024-03501-4

22.

Kim

S. D.

Allen

N. E.

Canning

C. G.

Fung

V. S. C.

(2013). Postural instability in patients with Parkinson’s disease. CNS Drugs, 27(2), 97–112. https://doi.org/10.1007/s40263-012-0012-3

23.

Lei

Sunzi

Dai

Liu

Wang

Zhang

(2019). Effects of virtual reality rehabilitation training on gait and balance in patients with Parkinson’s disease: A systematic review. PloS One, 14(11), e0224819. https://doi.org/10.1371/journal.pone.0224819

24.

Liang

S. G.

Chow

J. C. M.

Leung

N. M.

Y. N.

T. M. H.

Woo

C. L. C.

Lam

F. M. H.

(2025). The effects of ankle and foot exercises on ankle strength, balance, and falls in older people: A systematic review and meta-analysis. Physical Therapy, 105(1), pzae157. https://doi.org/10.1093/ptj/pzae157

25.

Liang

S. G.-S.

Fan

E. S.-L.

Lam

P. K.

Kwok

W. T.

C. Z.-H.

Lam

F. M.-H.

(2024). The effect of adding real-time postural feedback in balance and mobility training in older adults: A systematic review and meta-analysis. Archives of Gerontology and Geriatrics, 123(1), 105439. https://doi.org/10.1016/j.archger.2024.105439

26.

Maddumage Dona

S. P.

Chand

P. B.

Haden

Kalyani

Lehn

Sullivan

Kerr

G. K.

(2025). Effectiveness of haptic cues on gait in people with Parkinson’s disease - A systematic review and meta-analysis. Gait and Posture, 122, 358–374. https://doi.org/10.1016/j.gaitpost.2025.08.069

27.

Mak

M. K.

Wong-Yu

I. S.

Shen

Chung

C. L.

(2017). Long-term effects of exercise and physical therapy in people with Parkinson disease. Nature Reviews: Neurology, 13(11), 689–703. https://doi.org/10.1038/nrneurol.2017.128

28.

Maxwell

J. P.

Masters

R. S. W.

Kerr

Weedon

(2001). The implicit benefit of learning without errors. The Quarterly Journal of Experimental Psychology A, 54(4), 1049–1068. https://doi.org/10.1080/02724980143000073

29.

Mayo

N. E.

Mate

K. K. V.

Fellows

L. K.

Morais

J. A.

Sharp

Lafontaine

A. L.

Hill

E. T.

Dawes

Sharkh

A. A.

(2024). Real-time auditory feedback for improving gait and walking in people with Parkinson’s disease: A pilot and feasibility trial. Pilot Feasibility Stud, 10(1), 115. https://doi.org/10.1186/s40814-024-01542-z

30.

McMaster

Cole

M. H.

Chalkley

Creaby

M. W.

(2022). Gait biofeedback training in people with Parkinson’s disease: A pilot study. Journal of Neuroengineering and Rehabilitation, 19(1), 72. https://doi.org/10.1186/s12984-022-01051-1

31.

Nachev

Kennard

Husain

(2008). Functional role of the supplementary and pre-supplementary motor areas. Nature Reviews Neuroscience, 9(11), 856–869. https://doi.org/10.1038/nrn2478

32.

Nanhoe-Mahabier

Allum

J. H.

Pasman

E. P.

Overeem

Bloem

B. R.

(2012). The effects of vibrotactile biofeedback training on trunk sway in Parkinson’s disease patients. Parkinsonism & Related Disorders, 18(9), 1017–1021. https://doi.org/10.1016/j.parkreldis.2012.05.018

33.

NHLBI . (2014). Quality assessment tool for before-after (pre-post) studies with no control group. Retrieved 30 August 2025 from: https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools

34.

Orrell

A. J.

Eves

F. F.

Masters

R. S.

(2006a). Motor learning of a dynamic balancing task after stroke: Implicit implications for stroke rehabilitation. Physical Therapy, 86(3), 369–380. https://doi.org/10.1093/ptj/86.3.369

35.

Orrell

A. J.

Eves

F. F.

Masters

R. S. W.

(2006b). Implicit motor learning of a balancing task. Gait and Posture, 23(1), 9–16. https://doi.org/10.1016/j.gaitpost.2004.11.010

36.

Pan

Tang

Duan

Nong

Wang

(2021). Global trends in the incidence, prevalence, and years lived with disability of Parkinson’s disease in 204 countries/territories from 1990 to 2019. Frontiers in Public Health, 9(1), 776847. https://doi.org/10.3389/fpubh.2021.776847

37.

Page

M. J.

McKenzie

J. E.

Bossuyt

P. M.

Boutron

Hoffmann

T. C.

Mulrow

C. D.

Shamseer

Tetzlaff

J. M.

Akl

E. A.

Brennan

S. E.

Chou

Glanville

Grimshaw

J. M.

Hróbjartsson

Lalu

M. M.

Loder

E. W.

Mayo-Wilson

McDonald

Moher

(2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. Journal of Clinical Epidemiology, 134(1), 178–189. https://doi.org/10.1016/j.jclinepi.2021.03.001

38.

Pallavi

Raghuvanshi

Kumar

S. D.

Patel

Kanetkar

Chhatlani

Rana

Betai

Rajan

Lahiri

(2025). On-demand cueing sensitive to step variability: Understanding its impact on gait of individuals with Parkinson’s disease. IEEE Journal of Translational Engineering in Health and Medicine, 13(1), 183–192. https://doi.org/10.1109/jtehm.2025.3563381

39.

Palmisano

Kullmann

Hanafi

Verrecchia

Latoschik

M. E.

Canessa

Fischbach

Isaias

I. U.

(2022). A fully-immersive virtual reality setup to study gait modulation. Frontiers in Human Neuroscience, 16(1), 783452. https://doi.org/10.3389/fnhum.2022.783452

40.

Pompeu

J. E.

Mendes

F. A.

Silva

K. G.

Lobo

A. M.

Oliveira Tde

Zomignani

A. P.

Piemonte

M. E.

(2012). Effect of nintendo Wii-based motor and cognitive training on activities of daily living in patients with Parkinson’s disease: A randomised clinical trial. Physiotherapy, 98(3), 196–204. https://doi.org/10.1016/j.physio.2012.06.004

41.

Radder

D. L. M.

Lígia Silva de Lima

Domingos

Keus

S. H. J.

van Nimwegen

Bloem

B. R.

de Vries

N. M.

(2020). Physiotherapy in Parkinson’s disease: A meta-analysis of present treatment modalities. Neurorehabilitation and Neural Repair, 34(10), 871–880. https://doi.org/10.1177/1545968320952799

42.

Salamh

P. A.

Kolber

M. J.

(2025). The anatomy of a systematic review: Does it meet the minimum threshold for review? Physiotherapy Theory and Practice, 41(6), 1133–1134. https://doi.org/10.1080/09593985.2025.2496843

43.

Santos

Scaldaferri

Santos

Ribeiro

Neto

Melo

(2019). Effects of the Nintendo Wii training on balance rehabilitation and quality of life of patients with Parkinson’s disease: A systematic review and meta-analysis. NeuroRehabilitation, 44(4), 569–577. https://doi.org/10.3233/NRE-192700

44.

Santos

S. M.

da Silva

R. A.

Terra

M. B.

Almeida

I. A.

de Melo

L. B.

Ferraz

H. B.

(2017). Balance versus resistance training on postural control in patients with Parkinson's disease: A randomized controlled trial. European Journal of Physical and Rehabilitation Medicine, 53(2), 173–183. https://doi.org/10.23736/s1973-9087.16.04313-6

45.

Smania

Corato

Tinazzi

Stanzani

Fiaschi

Girardi

Gandolfi

(2010). Effect of balance training on postural instability in patients with idiopathic Parkinson’s disease. Neurorehabilitation and Neural Repair, 24(9), 826–834. https://doi.org/10.1177/1545968310376057

46.

Wilson

Alcock

Yarnall

A. J.

Lord

Lawson

R. A.

Morris

Taylor

J.-P.

Burn

D. J.

Rochester

Galna

(2020). Gait progression over 6 years in Parkinson’s disease: Effects of age, medication, and pathology. Frontiers in Aging Neuroscience, 12(1), 577435. https://doi.org/10.3389/fnagi.2020.577435

47.

Winkler

DeMarch

Campbell

Smith

(2022). Use of real-time multimodal sensory feedback home program improved backward stride and retention for people with Parkinson disease: A pilot study. Clinical Parkinsonism & Related Disorders, 6(1), 100132. https://doi.org/10.1016/j.prdoa.2022.100132

48.

Yang

W. C.

Wang

H. K.

R. M.

C. S.

Lin

K. H.

(2016). Home-based virtual reality balance training and conventional balance training in Parkinson’s disease: A randomized controlled trial. Journal of the Formosan Medical Association, 115(9), 734–743. https://doi.org/10.1016/j.jfma.2015.07.012

49.

Zadeh

A. K.

Sadeghbeigi

Safakheil

Setarehdan

S. K.

Alibiglou

(2024). Connecting the dots: Sensory cueing enhances functional connectivity between pre-motor and supplementary motor areas in Parkinson’s disease. European Journal of Neuroscience, 60(3), 4332–4345. https://doi.org/10.1111/ejn.16437