Comparing rehabilitation with and without neuromuscular electrical stimulation in hospitalized older adults: Cognitive, functional, and acceptability outcomes from a pragmatic pilot study

Abstract

Neuromuscular electrical stimulation (NMES) is a non-invasive therapeutic approach offering targeted muscle contraction without requiring voluntary effort. This pragmatic pilot study investigated the effects of integrating NMES into usual rehabilitation care on cognitive and mobility outcomes, as well as its acceptance among hospitalized older adults in rehabilitation settings. Twenty-nine older adults with a mean age of 81.7 years (SD = 9.0) participated in this study. Patients received either usual rehabilitation care alone (5 days/week) or NMES in addition to usual rehabilitation care (3 days/week) for three weeks. Cognitive function and mobility were assessed pre- and post-intervention, and NMES acceptability was measured using an 8-item questionnaire. Integrating NMES to usual rehabilitation care significantly improved cognitive flexibility, demonstrated by reduced task-switching errors (p = 0.03). No additional benefits were observed in other cognitive or mobility measures, although all participants showed mobility improvements over time. NMES was considered acceptable by older adults in the rehabilitation setting.Integrating NMES into usual rehabilitation is well accepted by hospitalized older adults, with preliminary evidence suggesting it may improve cognitive flexibility.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s11357-025-02030-9.

Introduction

Physical exercise is one of the most effective non-pharmacological strategies to prevent and manage the decline in functional capacity in older adults. Its benefits extend to key domains such as mobility and cognition and have been consistently supported by meta-analytic evidence over the past decade [1–3]. However, maintaining regular physical exercise still remains a significant challenge for many older adults. Beyond motivational factors, barriers such as mobility restrictions, pain, and hospitalization hinder the sustainability of physical activity and exercise programs [4, 5]. Medical care and rehabilitation services, for instance, are such settings where physical activity is often significantly limited due to these factors. Given these activity-limiting barriers, neuromuscular electrical stimulation (NMES) may emerge as an adjunct or alternative strategy by eliciting muscle contractions through electrical impulses, thereby mimicking the physiological effects of exercise. NMES utilizes transcutaneous electrical currents to stimulate muscle fibers to contract and does not require voluntary effort and active cooperation by the participant [6]. Therefore, it can be applied at the patient's bedside, making it suitable for individuals with limited mobility and little or no voluntary participation, thus enabling therapeutic engagement. Considering older adults represent a substantial proportion of patients in rehabilitation settings, often experiencing post-acute injuries and age-related deconditioning, integrating NMES into geriatric rehabilitation may offer a valuable strategy to support functional recovery and maintain independence. Indeed, NMES has not only been shown to improve mobility and overall functional performance in healthy older adults [7–10], but also to preserve muscle function in critically ill and frail patients, including those with acute myocardial infarction and postoperative arthroplasty by maintaining muscle strength and range of motion and enhancing gait speed and balance [11–14]. However, despite its potential, it has been reported that NMES still suffers from poor clinical acceptance in rehabilitation settings [15].

Beyond its peripheral effects, NMES may also influence cognitive function. Although growing evidence suggests a crosstalk between the skeletal muscle and the brain [16, 17], the potential effects of electrically induced muscle contractions on brain function have received comparatively less attention. Further, one notable advantage of NMES may be that it provides a model for studying muscle–brain crosstalk by directly stimulating muscles without requiring volitional effort or engaging the cardiovascular system to the same extent as whole-body exercise. This allows for a more controlled and isolated investigation of muscle-driven effects on brain function, minimizing the influence of psychological and motivational factors that typically accompany voluntary physical activity.

Nevertheless, the effects of NMES on cognitive function remain unclear, as the available literature is limited, methodologically heterogeneous, and reports conflicting findings. Most evidence comes from acute interventions: some studies have reported improvements [18, 19], whereas others found no significant changes in executive function, attention, or memory [20, 21]. The only chronic trial to date indicated that stroke patients receiving 48 sessions of NMES in addition to usual care demonstrated greater cognitive improvements than those receiving usual care alone [22]. Although research on NMES and cognition is sparse, existing findings allow hypotheses about possible underlying mechanisms. On the one hand, humoral changes may be involved, as NMES has been reported to increase circulating muscle-derived factors (myokines) such as brain-derived neurotrophic factor (BDNF), lactate, cathepsin B, insulin-like growth factor-1 (IGF-1), and interleukin-10, all of which have been implicated in synaptic plasticity and cognitive processes [23–26]. On the other hand, neural pathways may also play a role, since afferent signals generated by electrically induced muscle contractions are transmitted to cortical and subcortical structures, activating sensorimotor and associative networks that could indirectly support higher-order cognition [27–31]. While preliminary and heterogeneous, these pathways provide a biologically plausible rationale for investigating whether NMES might exert cognitive effects in older adults. However, cognitive outcomes of NMES remain largely underexplored in hospitalized older adults, and the acceptability of integrating NMES into routine geriatric rehabilitation has not been examined.

Given that NMES may mimic certain aspects of physical exercise [32] and addresses some of the key barriers older adults face in rehabilitation, the present study aimed to examine whether rehabilitation care including NMES could enhance cognitive and mobility outcomes of older adults. Additionally, since user acceptance is a key factor for the successful and widespread implementation of such interventions in rehabilitation settings, we also aimed to evaluate the acceptability of NMES among older adults. We hypothesized that rehabilitation with NMES would lead to greater improvements in cognition and mobility compared to rehabilitation without NMES, and that it would be considered acceptable by the participants.

Methods

Study design and participants

This study was designed as a pragmatic pilot trial, that is, a small-scale exploratory study embedded in routine hospital rehabilitation, with the specific interest of assessing acceptability and preliminary cognitive/functional signals under real-world conditions in order to inform the design of a future definitive trial [33]. Consistent with the CONSORT 2010 extension for pilot and feasibility trials, the sample size was not based on a formal power calculation. Instead, it was determined based on practical considerations, including the recruitment period, unit capacity, and expected patient flow [34]. The two-arm pragmatic comparative study was conducted at the Medical and Rehabilitation Care (MCR) unit of the University Hospital of Caen between January and June 2024. Twenty-nine patients (mean age = 81.7 ± 9.0 years) admitted to the unit were allocated by rehabilitation staff to either usual rehabilitation care alone or usual care including NMES. Allocation followed routine unit logistics (admission timing and slot availability) without investigator discretion; when both options were simultaneously available, a simple randomization list generated with GraphPad QuickCalcs was used to ensure balanced distribution. This procedure is better described as a pragmatic allocation process, effectively random in nature, and consistent with the aims of a pilot pragmatic trial. Inclusion criteria were: (i) being 65 years or older, (ii) being a patient admitted to a Medical Care and Rehabilitation (MCR) service, (iii) ability to communicate fluently in French, and (iv) having read the information sheet and provided written informed consent to participate in the study. Exclusion criteria were: (i) unable to walk 10 m (use of a walking aid such as a cane was permitted), (ii) having an implanted cardiac pacemaker or defibrillator, or (iii) having a known diagnosis of epilepsy. The eligibility criteria were verified by the medical team. No compensation was provided, as all care and assessments conducted were part of the routine management of patients in the MCR unit. After a minimum two-day reflection period, patients provided written informed consent. The study was approved by the Comité d’Ethique pour la Recherche en Sciences et Techniques des Activités Physiques et Sportives (CERSTAPS) (ethics no: IRB00012476_2024_02_02_293). Patient characteristics are presented in Table 1.

Table 1.

Patient characteristics

	NMES	Usual care
Age (years)	80.9 (9.3)	82.6 (8.9)
Sex (F/M)	8/6	12/3
Weight (kg)	69.7 (14.7)	68.7 (16.0)
Height (cm)	162.8 (8.7)	164.2 (9.4)
BMI (kg/m2)	26.2 (4.9)	24.7 (5.2)
Education (years)	8.0 (5.2)	11.7 (4.4)
Hand preference
Right-handed	11	12
Left-handed	3	1
Ambidextrous	0	2
Comorbidities (n)	1.7 (1.2)	1.8 (1.0)
Abdominal aortic aneurysm	-	1
Atrial fibrillation	-	1
Breast cancer	1	-
Chronic pain	2	3
Chronic obstructive lung disease	1	-
Cerebral vascular accident	-	1
Hearing loss	1	1
Hypertension	8	8
Hyperthyroidism	-	2
Macula degeneration	1	-
Osteoporosis	-	3
Peripheral obliterative arteriopathy	1	1
Sleep apnea syndrome	1	-
Transient ischemic attack	2	2
Type 2 diabetes	4	2
Vertigo	1	-
Visual impairment	1	-

Mean ± SD values are represented for numerical variables. The number of participants (n) is represented for categorical variables

Measures

Anthropometric measurements consisted of body mass (kg), height (cm), and body mass index (BMI), obtained with a portable scale and a stadiometer. Hand preference was evaluated via the Edinburgh Handedness Inventory [35]. Cognitive and physical outcome measures were assessed before and after the intervention period. Participants were tested at the same time of day (between 9:00 a.m. and 12:00 p.m.). Post-tests were conducted the day after the final intervention session, matching the timing of the pre-tests. Cognitive tests were administered before the physical assessments. To reduce order effects, the sequence of tests within each domain was varied across participants. All assessments were administered by a master’s student who was not involved in the data analysis.

Cognitive function

Cognitive performance was assessed using the Test of Attentional Performance (TAP 2.3.1) battery, specifically through three subtests: Working Memory, Flexibility, and Go/NoGo [36]. The tests were administered on a 24-inch laptop (1920 × 1080 pixels) in a quiet room, with participants seated approximately 70 cm from the screen. Participants who regularly used vision aids wore them during the testing. Each subtest lasted approximately 5 min and was preceded by practice trials lasting 1 to 2 min.

Task-switching

Cognitive flexibility was assessed using the letter–number task-switching test. In each trial, a letter and a number were simultaneously displayed on the left and right sides of the screen. Participants were instructed to alternate their responses between the side containing the letter and the side containing the number, beginning with the letter. They responded by pressing a left or right key corresponding to the side of the target stimulus. Errors and response times were recorded.

2-back test

Working memory was evaluated using a visual 2-back version of the n-back test. In this task, single-digit numbers were presented sequentially on a screen. Participants were instructed to press a key as quickly as possible whenever the current number matched the one presented two positions before in the sequence. Accuracy was recorded.

Go/NoGo

Inhibitory control was assessed using the Go/NoGo test. Five squares with different visual patterns appeared simultaneously on the screen in random combinations. Participants were instructed to press a key as quickly as possible only when two of the five squares matched the previously shown target patterns. Accuracy and response times were recorded.

Physical function

The Five Times Sit to Stand Test (5XSST)

Participants were seated on height-adjustable, armless chairs with their arms crossed over the chest (hands resting on opposite shoulders) and were instructed to stand up and sit down five times [37]. The total duration to complete five repetitions was recorded as the test score.

6-Minute Walk Test (6MWT)

Participants performed the test by walking back and forth continuously along a hallway between two cones placed 30.5 m apart. They were instructed to walk as much as possible in 6 min [38]. The total distance walked was recorded in meters.

Tinetti Performance-Oriented Mobility Assessment

Participants performed the test by completing balance and gait tasks, including sitting, standing, turning, and walking. Each component was scored separately, with a maximum of 16 points for balance and 12 points for gait. A total score out of 28 was recorded, with scores ≤ 18 considered a high risk of falling [39].

Timed Up and Go Test (TUG)

Timed Up and Go Test (TUG): The participant sat in a standard chair with their back against the backrest and arms on the chair’s arms, wearing regular footwear. A 3-m walkway was marked, and the patient was instructed to stand up on the word “Go,” walk the 3-m line at a comfortable pace, turn around, walk back, and sit down. Timing started on “Go” and stopped when the subject’s hip touched the seat. Time is recorded in seconds. For the dual-task (DT) condition, participants performed the same mobility task while simultaneously engaging in counting backward by five from 100. They were instructed not to stop walking even when they made a mistake and were requested to give equal attention to both tasks and not to prioritize any. DT cost was calculated by dividing the time difference between the TUG test and the TUG-DT by the completion time of the TUG, represented as a percentage. A positive value indicates a decline in dual-task performance [40].

Acceptability

Acceptability was evaluated at the end of the intervention using an 8-item questionnaire developed by Sekhon et al., based on the Theoretical Framework of Acceptability (TFA) [41, 42]. Each item corresponded to one of the seven constructs of the TFA: affective attitude, burden, ethicality, perceived effectiveness, intervention coherence, self-efficacy, and opportunity costs, with an additional item assessing general acceptability. Participants rated each item on a 5-point Likert scale. Acceptability was determined by calculating the proportion of responses rated ≥ 3, which was considered satisfactory [43]. In accordance with Sekhon et al.’s approach, the general acceptability item was treated as a validation item and not included in the composite acceptability score calculation.

Intervention

Following enrollment and as soon as possible after providing informed consent, patients started receiving treatment. All patients received the usual care, while the NMES group also underwent additional NMES sessions as part of the therapeutic approach.

Usual care

All patients received standard care as part of their rehabilitation program, including active-assisted and/or active exercises, depending on each patient’s physical capacity and level of compliance. The usual care followed the adapted physical activity approach and consisted of stretching, range of motion exercises, muscle strengthening, balance training, and aerobic exercises. Sessions were conducted individually and lasted approximately 40 min, five days per week for three weeks. Usual care rehabilitation was delivered by physical therapists routinely working in the rehabilitation unit.

NMES

Patients received additional NMES sessions targeting the knee extensors and plantar flexor muscles bilaterally, as part of their standard rehabilitation program, using the Compex SP 2.0 device (Compex Medical SA, Geneva, Switzerland). Stimulation was applied using surface electrodes (5 × 5 cm, axion® GmBh, Leonberg, Germany), with electrode placement and device setup performed by a trained physical therapist. All NMES sessions were delivered by this single therapist to ensure consistency of the stimulation protocol. The therapist also monitored the skin for irritation during the initial sessions. Each NMES session included a 5-min warm-up at 5 Hz, followed by a main stimulation phase at 70 Hz consisting of 52 cycles with 8 s of stimulation followed by 4 s of rest, and concluded with a 10-min recovery period at 3 Hz. The pulse width was set to 300 μs. The current intensity was gradually increased based on the patient's maximum tolerable level, ensuring visible muscle contractions. This stimulation protocol was selected based on current evidence supporting its functional efficacy in older adults [44]. The NMES intervention started on the day of enrollment, following baseline assessments, and was delivered thrice weekly for three weeks. To minimize interference with routine rehabilitation care, NMES sessions were systematically scheduled at different times from usual care sessions, with morning therapy paired with afternoon NMES (and vice versa), ensuring temporal separation between interventions.

Statistical analysis

Linear mixed-effects models (LMEMs) were employed to assess the effect of the intervention on cognitive performance and mobility over time, using the lme4 and emmeans packages in R. These models accounted for both fixed effects (Time, group, and their interaction) and random effects (subject ID), allowing for individual variability. For mobility-related outcomes, models were adjusted for age, sex, and BMI. For cognitive outcomes, adjustments were made for age, sex, and education level. All models were fitted using maximum likelihood (ML) estimation, which uses all available data under the assumption of data missing at random (MAR) and thereby accounts for missing observations without the need for imputation. Acceptability scores were compared between intervention groups using independent samples t-tests. All statistical analyses were performed using RStudio (Version 2024.12.0 + 467), corresponding analysis code is available on GitHub at: https://github.com/evrimgokce/NMES.

Results

Of the 29 patients enrolled, 28 completed the intervention, one patient died during hospitalization, from causes unrelated to the study. Those who completed the intervention attended 100% of their scheduled sessions, and no adverse events related to NMES or usual rehabilitation care were observed. The primary diagnoses of patients were as follows (n, NMES/APA): fractures (n = 7, 3/4), total hip arthroplasty (n = 3, 2/1), stroke (n = 2, 1/1), myocardial infarction (n = 2, 1/1), post-surgery (n = 4, 2/2), gastrointestinal hemorrhage (n = 1, 0/1), non-fracture falls (n = 5, 2/3), and generalized deconditioning or muscle weakness (n = 5, 2/3).

Supplementary Fig. S1 shows the CONSORT diagram. Table 2 presents descriptive statistics for cognitive and mobility outcomes. Table 3 reports the main fixed effects, and full model results including covariates are available in Supplementary Table S1.

Table 2.

Descriptive statistics for cognitive and mobility outcomes at baseline and post-intervention

Outcome	NMES				Usual care
	Pre	n	Post	n	Pre	n	Post	n
Cognitive
TS Error	8.90 (11.19)	10	3.78 (7.01)	9	5.73 (5.53)	11	6.60 (6.54)	10
TS RT	2102.88 (1364.62)	10	4359.57 (6916.91)	9	2261.18 (1799.14)	11	1718.12 (587.71)	10
2-back test Acc	8.89 (3.82)	9	11.22 (2.59)	9	10.18 (3.84)	11	10.00 (3.86)	12
Go/NoGo Acc	18.17 (5.27)	12	18.11 (7.46)	9	17.83 (6.26)	12	21.10 (9.06)	10
Go/NoGo RT	960.17 (306.07)	12	928.12 (477.76)	9	1015.68 (336.14)	12	1054.31 (568.59)	10
Mobility
5XSST	18.26 (8.75)	11	16.98 (5.78)	13	24.64 (7.99)	11	21.12 (8.38)	14
6MWT	126.92 (100.51)	13	199.23 (112.97)	13	87.47 (58.72)	15	182.00 (110.92)	15
TUG-ST	31.29 (25.77)	12	22.29 (15.84)	12	32.87 (15.67)	12	31.32 (22.46)	14
TUG-DT	42.58 (35.90)	12	26.70 (18.89)	10	34.90 (12.78)	9	30.39 (20.98)	12
Tinetti test	19.86 (6.92)	14	23.54 (4.39)	13	19.27 (6.36)	15	23.00 (5.53)	15

5xSST Five Times Sit-to-Stand Test, 6MWT Six-Minute Walking Test, Acc Accuracy, Chair test DT, Dual Task, RT Reaction Times, ST Single Task, TS Task-Switching, TUG Timed Up and Go Test

Results are presented as mean (SD). n indicates the number of participants with available data at each time point.

Table 3.

Linear mixed-effect model outputs for cognitive and mobility outcomes

	Estimate	SE	95% CI	t	p
TS-Error
Time	−5.74	2.33	−10.31 – −1.16	−2.45	0.022*
Group	−4.64	3.32	−11.16 – 1.87	−1.39	0.172
Time*Group	6.72	3.17	0.48–12.94	2.11	0.047*
TS-RT
Time	167.64	180.01	−185.17 – 520.45	0.93	0.365
Group	−2617.71	1856.89	−6257.15 – 1021.72	−1.41	0.172
Time*Group	−152.4	240.1	−622.98 – 318.17	−0.63	0.534
2-back test Acc
Time	1.39	1.32	−1.19 – 3.98	1.05	0.303
Group	0.61	1.62	−2.56 – 3.79	0.37	0.708
Time*Group	−1.07	1.72	−4.46 – 2.31	−0.62	0.54
Go/NoGo Acc
Time	−0.55	2.11	−4.69 – 3.58	−0.26	0.794
Group	0.06	2.86	−5.55 – 5.67	0.02	0.983
Time*Group	3.39	2.89	−2.27 – 9.07	1.17	0.253
Go/NoGo RT
Time	−21.63	132.03	−280.42 – 237.15	−0.16	0.871
Group	41.44	175.31	−302.15 – 385.05	0.23	0.814
Time*Group	92.55	181.1	−262.41 – 447.52	0.51	0.615
5XSST
Time	−3.43	1.63	−6.64 – −0.23	−2.1	0.048*
Group	4.77	2.92	−0.95 – 10.49	1.63	0.11
Time*Group	−1.45	2.32	−6.01 – 3.10	−0.62	0.539
6MWT
Time	72.3	21.74	29.69–114.91	3.32	0.002*
Group	−22.44	32.6	−86.34 – 41.45	−0.68	0.495
Time*Group	33.3	30.74	−26.96 – 93.56	1.08	0.288
Tinetti test
Time	3.81	1.41	1.05–6.58	2.7	0.011*
Group	0.28	2.07	−3.79 – 4.35	0.13	0.892
Time*Group	0.79	2	−3.12 – 4.72	0.39	0.693
TUG-ST
Time	−10.22	4.96	−19.94 – −0.49	−2.05	0.051
Group	2.5	7.89	−12.96 – 17.97	0.31	0.752
Time*Group	4.49	7.02	−9.27 – 18.27	0.64	0.528
TUG-DT
Time	−18.73	6.76	−31.99 – −5.48	−2.77	0.011*
Group	−8.5	9.78	−27.67 – 10.66	−0.87	0.389
Time*Group	12.88	9.79	−6.31 – 32.09	1.31	0.202

SE standard errors, 95% CI 95% confidence interval, 5xSST Five Times Sit-to-Stand Test, 6MWT Six-Minute Walking Test, Acc Accuracy, Chair test DT, Dual Task, RT Reaction Times, ST Single Task, TS Task-Switching, TUG Timed Up and Go Test

p values were calculated from LMEMs

* indicates statistical significance at p < 0.05

Cognitive outcomes

A linear mixed-effects model revealed a Time × Group interaction effect for task-switching error score, β = 6.72, SE = 3.18, t(20.38) = 2.11, p = 0 0.04. Follow-up contrasts indicated that the NMES group showed a significant reduction in error rates from pre- to post-intervention (p = 0 0.03), suggesting improved cognitive flexibility following the intervention (Fig. 1).

Fig. 1.

Task-switching error scores before and after the intervention

No other cognitive outcome exhibited a significant time × group interaction. Full model results are presented in Table 3.

Mobility outcomes

No significant Time × Group interaction effects were observed for any of the mobility outcomes, including the 6MWT, TUG (single and dual-task), 5XSST and Tinetti Test.

There was a main effect of time for 6MWT (b = 72.31, p = 0.002), 5XSST (b =  − 3.43, p = 0.048), TUG-DT (b =  − 18.73, p = 0.011), and Tinetti Test (b = 3.81, p = 0.011) indicating overall improvement across participants, independent of intervention assignment (Fig. 2). Specifically, participants demonstrated a significant increase in the distance walked in 6MWT, completed the 5XSST and TUG-DT tasks faster, and scored higher for the Tinetti test. Full model results are presented in Supplementary Table S1.

Fig. 2.

Changes in mobility outcomes over time

Acceptability

There was no significant difference in overall acceptability scores between groups, t(21.83) = 1.45, p = 0.161, 95% CI [− 0.108, 0.610], although the NMES group had a slightly higher mean score (M = 3.81) than the usual care group (M = 3.56). Similarly, “general acceptability” score was slightly higher in the NMES group (M = 4.30) than usual care (M = 4.13), but the difference was not significant. A marginal effect was observed for perceived effectiveness (p = 0.052), with higher ratings in the NMES group (M = 4.15) compared to usual care (M = 3.46) (Fig. 3).

Fig. 3.

Individual item scores for acceptability questionnaire by intervention

Discussion

The present pilot study investigated the effects of adding NMES to usual care on cognitive and physical outcomes, and its acceptability among hospitalized older adults in a rehabilitation setting. Participants receiving rehabilitation including NMES, compared to those not receiving NMES, demonstrated significant improvement in cognitive flexibility, as evidenced by reduced task-switching errors. While there were no significant differences in mobility outcomes between the groups, both showed improvement over time. Additionally, the rehabilitation involving NMES was rated as acceptable by older adults, with acceptability scores comparable to rehabilitation without NMES, and slightly higher ratings for perceived effectiveness.

While our study demonstrated benefits specifically in cognitive flexibility, prior work with chronic NMES in stroke patients reported improvements in broader measures of cognitive function [22]. Cognitive flexibility is a critical executive function that enables individuals to adapt their behavior to changing or complex environmental demands, and it tends to decline with age [45]. Research has shown that poor cognitive flexibility is strongly associated with increased fall risk in older adults [46]. Our preliminary finding suggests that NMES may have a positive impact on executive functioning in older adults. This approach could be particularly helpful for older adults who may face challenges or reluctance to engage in cognitively or physically demanding activities, such as cognitive training or structured exercise. However, these results need to be confirmed in larger, controlled studies before any clinical implications can be established.

Although we assessed multiple executive domains, only cognitive flexibility improved in our sample. Prior research suggests that, as people age, cognitive flexibility retains domain-specific variance, whereas inhibition and working memory reflect more shared executive function processes [47]. This structural distinction does not imply that NMES selectively targets flexibility; however, flexibility tasks may be more sensitive to detecting early or subtle changes in executive functioning in small pilot samples or short interventions. Given the pilot nature of our study, however, this finding should be interpreted with caution, and firm conclusions about differential responsiveness to NMES should be avoided.

The mechanisms often proposed to explain exercise-induced cognitive enhancement, such as the activation of muscle-derived circulating signaling pathways and increased neuronal activity-dependent neuroplasticity [48, 49], may also be applicable to NMES, suggesting its potential to elicit similar cognitive benefits in older adults.

For instance, brain-derived neurotrophic factor (BDNF), a well-known exercise-induced myokine linked to cognitive improvement, has also been shown to increase after NMES, suggesting a potential mechanism for its cognitive effects [23, 24]. Furthermore, studies have also reported increased levels of myokines following NMES, including lactate, cathepsin B, IGF-1, and interleukin-10, which are also known to be associated with cognitive function [18, 25, 26].

In addition to myokine-mediated signaling, the neuronal pathway offers another potential mechanism through which NMES may influence cognition. Contrary to general belief, NMES does not bypass the central nervous system due to the afferent input from muscle spindles, joints, and skin that travels to the cortex. These afferent inputs engage not only the motor system but also broader neural circuits involved in sensory and cognitive processing [27]. Supporting this, neuroimaging studies using fNIRS, fMRI, and EEG have demonstrated that NMES activates sensorimotor networks and subcortical structures [28–30], suggesting the potential for rapid plastic changes within the sensorimotor regions [31]. These changes may involve the activation of previously silent synaptic connections, contributing to longer-term neuroplastic effects [50]. Given that the sensorimotor system is increasingly recognized as contributing to cognitive domains such as attention, perception, executive function, and language [51, 52], NMES-induced neural effects may support cognitive improvement.

Importantly, the extent to which NMES engages the central neural pathway appears to depend on specific stimulation parameters. Using wider pulse widths, higher frequencies, and greater intensities can recruit central pathways through sensory axons, potentially leading to a greater neural impact [30, 53, 54]. Consequently, these insights highlight the need for parameter-specific design when targeting cognitive outcomes with NMES.

Our findings regarding mobility outcomes indicated that adding NMES to usual care did not provide any additional benefits compared to usual care alone. While some studies have demonstrated positive effects of NMES on mobility in older adults, many lacked control groups [7–9]. On the other hand, a recent pilot study showed improved functional mobility in older adults receiving a 12-week NMES program alongside usual care, while another study of older patients with acute myocardial infarction reported increased lower limb strength after seven NMES sessions [10, 14]. However, the contexts of these studies differ from ours. The 12-week pilot study involved functionally fit nursing home residents, potentially more responsive to NMES, whereas the MI study applied NMES alongside a basic postoperative care program with limited exercise rehabilitation, which may have left more room for improvement. In contrast, our participants were older adults with functional impairments undergoing an adapted physical activity. Therefore, one possible explanation for the lack of additional NMES effect might be that the usual care provided was already sufficiently effective in improving mobility, limiting the potential for further gains.

In addition, variability in NMES parameters may have also contributed to inconsistent findings, suggesting that a one-size-fits-all approach may not be suitable. Literature generally indicates that higher intensity, frequency, and pulse width generally lead to greater muscle strength gains, with stimulation at the highest tolerable level recommended due to its linear relationship with force production [15, 55]. Lower frequencies tend to recruit slow-twitch, fatigue-resistant fibers, while higher frequencies target fast-twitch fibers linked to greater force [30, 56]. Thus, parameter selection also appears to be associated with the targeted muscle group and function, making generalizations more difficult. For instance, Glaviano et al. have demonstrated that a pulse width of 400–600 µs and frequency between 30–50 Hz provides optimal muscle recruitment without inducing muscle fatigue in the quadriceps [57]. Besides, to minimize the risk of muscle fatigue and allow muscle recovery, sufficient off-time between stimulation periods, such as a 1:5 ratio (e.g., 8 s on, 40 s off) has been recommended [55]. Given these considerations, certain aspects of our NMES protocol may have limited its effectiveness in promoting muscle strength. However, due to the lack of precise guidelines for tailoring NMES parameters to specific targets, designing an optimal protocol remains particularly challenging.

It is worth noting that the limited number of NMES sessions in our study (9 sessions) may have been insufficient to induce significant muscle strength gains. Previous studies often use more frequent and prolonged protocols, such as twice-daily sessions over several weeks, particularly for mobility-related outcomes [7–10, 22].

A key strength of this pilot trial is the demonstrated acceptability of NMES among hospitalized older adults, based on a structured TFA-based questionnaire. This finding aligns with existing literature showing NMES is comfortable for patients with specific medical conditions [58, 59]. Importantly, evidence from prior studies has shown that portable NMES devices can be safely implemented in home-based programs [60, 61]. Coupled with our finding of acceptability in older adults, this highlights the potential for NMES to be delivered beyond hospital settings, thereby reducing reliance on inpatient services and easing the care burden in aging populations. Given that NMES is a feasible and relatively low-cost intervention, future research should investigate optimized protocols and stimulation parameters, and assess their effects across different functional domains in both healthy older adults and those with specific pathologies. Moreover, given evidence that voluntary cognitive engagement can affect both physical and cognitive outcomes [62], future studies should consider its inclusion as an active control condition or as a complementary intervention alongside NMES.

This study is a pragmatic pilot trial embedded in routine rehabilitation, and provides preliminary signals rather than the causal certainty that a randomized controlled trial can evidence. These findings should therefore be confirmed in a fully powered, rigorously controlled trial before any conclusions regarding clinical implementation can be drawn. In addition to these considerations, several limitations should be acknowledged. First, the lack of standardized baseline screening for global cognition (e.g., Mini-Mental State Examination) and emotional status limited the baseline characterization of our sample, and undetected cognitive impairment may have introduced heterogeneity in cognitive outcomes. Second, medication data were not systematically collected, despite the fact that many older adults are polymedicated and certain drug classes may directly affect cognitive or physical performance. Third, the small number of intervention sessions due to the limited duration of the hospitalization period may have impacted the outcomes. Finally, the potential placebo effect of NMES was not controlled. Future studies should therefore incorporate standardized baseline cognitive and emotional screening, systematically document medication use, increase the number of intervention sessions, and include appropriate placebo or active control conditions to more rigorously evaluate the effectiveness of NMES.

Conclusion

Integrating NMES into usual rehabilitation appears feasible and acceptable for older adults in rehabilitation settings, with early results suggesting potential benefits for cognitive flexibility. A fully powered trial with target-specific NMES protocols is warranted to determine its efficacy as a clinical rehabilitation tool for functional mobility and cognition.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

Conceptualization, methodology, and writing original manuscript: EG; validation: AG, AL, GL; investigation and data collection: AH and EG; data curation and visualization: EG; supervision: AG, AL, GL. All authors have read and approved the final version of the manuscript.

Funding

AH was funded by the COMETE laboratory on internal funds. EG was co-funded by the Normandy County Council and the European Union in the framework of the ERDF-ESF operational programme 2021–2027 (#00159760).

Data availability

The datasets generated and analyzed during the current study are publicly available in the GitHub repository at https://github.com/evrimgokce/NMES.

Declarations

Competing interests

The authors declare no competing interests.