Effects of exercise-based rehabilitation training in schizophrenia patients

Abstract

BACKGROUND

Conventional antipsychotic treatment combined with health education shows limited efficacy in patients with schizophrenia, highlighting the need for further optimization of therapeutic strategies.

AIM

To evaluate the therapeutic efficacy of exercise-based rehabilitation training in patients with schizophrenia.

METHODS

From February 2022 to February 2025, 127 patients with schizophrenia were divided into a control group (n = 59) managed with routine pharmacotherapy and an intervention group (n = 68) receiving additional exercise-based rehabilitation training. Outcomes were assessed before and after the intervention using the Positive and Negative Syndrome Scale, the Nurses’ Observation Scale for Inpatient Evaluation (NOSIE), the Stroop Color-Word Test (SCWT), the Trail Making Test, the Brief Assessment of Cognition in Schizophrenia symbol coding test, the Inpatient Psychiatric Rehabilitation Outcome Scale, and serum brain-derived neurotrophic factor levels.

RESULTS

After the intervention, the intervention group demonstrated significantly greater reductions in Positive and Negative Syndrome Scale scores, NOSIE negative factor scores, Trail Making Test completion time, and Inpatient Psychiatric Rehabilitation Outcome Scale total and domain scores compared with the control group. Significant post-treatment increases were observed in NOSIE positive and total scores, SCWT word and color accuracy, Brief Assessment of Cognition in Schizophrenia performance, and serum brain-derived neurotrophic factor levels in the intervention group relative to baseline and control group values. SCWT total correct responses did not differ between groups or timepoints.

CONCLUSION

Exercise-based rehabilitation training significantly improves psychiatric symptoms, cognitive performance, social functioning, and biological markers in patients with schizophrenia, supporting its incorporation as an adjunctive component of standard treatment.

Key Words: Exercise-based rehabilitation training; Schizophrenia; Intervention effect; Cognitive performance; Social functioning

Core Tip: This study evaluated exercise-based rehabilitation training in 127 patients with schizophrenia using a comprehensive program integrating motor skills, dynamic cognitive challenges, and structured social interaction. The intervention demonstrated clear advantages in alleviating clinical symptoms and improving cognitive function, information processing speed, and social functioning, supporting its value as an effective adjunctive treatment option.

INTRODUCTION

Schizophrenia is a complex psychiatric syndrome characterized by abnormalities in mental and motor functioning. It is highly heritable, clinically heterogenous, and typically follows a chronic and relapsing course. Affected individuals display varying degrees of psychiatric symptoms, negative manifestations, cognitive impairments, affective dysregulation, and movement disorders[1]. These deficits impair information processing, attention, working memory, language learning and recall, visuospatial abilities, reasoning, and social cognition, ultimately leading to marked functional disability and occupational instability[2]. Epidemiological data indicate that the prevalence of schizophrenia has gradually increased in countries with a higher Social Demographic Index. Men tend to experience an earlier age at onset and bear a greater disease-related social burden[3]. The neurobiological basis of schizophrenia involves dysregulation of multiple neurotransmitter systems across distinct brain regions, including hyperactivity of glutamate/dopamine signaling in the basal ganglia and hypoactivity of dopamine, glutamate, and gamma-aminobutyric acid in cortical areas, particularly the frontal cortex[4]. Currently, pharmacological treatments primarily rely on dopamine D2 receptor antagonism, yet their efficacy and tolerability leave much to be desired[5]. Evidence from a randomized controlled trial suggests that standard antipsychotic therapy combined with health education is often insufficient to achieve meaningful recovery, with many individuals continuing to experience persistent negative symptoms and cognitive deficits. Exercise-based rehabilitation training has emerged as a promising adjunctive intervention. Programs incorporating aerobic endurance training, flexibility, strength training, and balance exercises have been shown to alleviate overall symptoms, boost functioning, and sharpen cognitive abilities[6]. These benefits may be mediated via dual mechanisms: Brain-derived neurotrophic factor (BDNF)-mediated thermoregulation and anti-inflammatory effects, and enhanced hypothalamic-pituitary-adrenal axis activity via prefrontal-hippocampal circuit stimulation[7].

Building upon standard antipsychotic treatment and health education, this study investigates whether adjunctive exercise-based rehabilitation training improves schizophrenia outcomes. Given the limited evidence on this combined approach, our findings contribute important data to an underexplored area.

MATERIALS AND METHODS

Case selection

Participants were eligible if they met all the following criteria: A diagnosis of schizophrenia according to established criteria[8]; stable disease with multiple episodes; good tolerance to antipsychotic medication; cohabitation with at least one caregiver; protocol adherence capability; unimpaired cognition/communication; and complete records. The exclusion criteria included substance-induced psychosis, deaf-mute individuals, cognitive-impairing neurological conditions, administration of any non-antipsychotic drugs that may affect cognitive function (except for trihexyphenidyl or benzodiazepines), cardiopulmonary/musculoskeletal impairments, severe systemic comorbidities, fracture-related complications, alcohol dependence, or perinatal status.

A total of 127 eligible schizophrenia patients treated during a three-year observation window (February 2022 to February 2025) were retrospectively included. Grouping was based on the treatment plan that each patient received. Of these, 59 patients receiving standard pharmacotherapy and health education were allocated to the control group, while 68 patients receiving supplemental exercise-based rehabilitation training were assigned to the intervention group. Baseline demographic and clinical characteristics did not differ significantly between the two groups (P > 0.05). The flowchart for patient inclusion and exclusion is presented in Figure 1.

Figure 1 Flowchart of patient inclusion and exclusion.

Intervention methods

Patients in the control group were maintained on conventional antipsychotic medication paired with health education. Antipsychotic drug treatment: The patients were administered Risperidone Tablets (Whenzhou KeMiao Biological Technology Co., Ltd., KM11140437), starting at a dose of 1 mg once daily, one tablet at a time. One week later, the dose could be adjusted based on the patient’s condition and adverse reactions, increasing to 2-4 mg, once or twice daily, two tablets each time. The treatment was continued for 8 weeks. Clinical staff implemented the following components: (1) Vigilant monitoring of psychiatric symptoms and stratified safety surveillance, prioritizing high-risk cases for incident prevention; (2) Supervised medication to promote adherence and therapeutic alliance, coupled with structured education on relapse prevention; and (3) Scheduled rehabilitation activities including self-care skill development (personal hygiene, dressing) and group-based activities (radio calisthenics, musical engagement) delivered in 30-45 minutes sessions, 3-5 times per week.

In addition to the above interventions, patients in the intervention group received exercise-based rehabilitation training. Training sessions lasted 50 minutes and were conducted 5 times per week for 8 weeks (total of 40 sessions). All sessions were supervised by both nurses and physical therapists. Each session began with a 10-minute warm-up consisting of 5 minutes of alternating slow and brisk walking followed by 5 minutes of dynamic stretching. The remaining 40 minutes were dedicated to task-oriented exercises. During weeks 1 and 2, training focused on coordinated ball handling skills using volleyball passing and softball toss-and-catch exercises. Patients were organized into groups of three, positioned in a triangular formation with 3-meter spacing. A designated passer initiated randomized throws to the other two receivers. Each training block lasted 5 minutes, and was repeated for five cycles, with 1-minute rest intervals between cycles. This was followed by a 10-minute instructional session introducing core softball maneuvers. During weeks 3 and 4, training emphasized route memorization walking. Nurses established five 10-meter walking tracks. Patients were required to memorize the landmarks and to alternate between routes as directed within a 10-minute window, completing four full rounds. During weeks 5 through 8, training transitioned to integrated exercises involving sandbag and table tennis drills. Working in groups of three, each member alternated between throwing sandbags for two minutes and practicing evasion for one minute per rotation. This sequence repeated for ten rounds with 60-second rest intervals between rounds.

Data collection and outcome measures

Psychiatric symptom improvement assessment: Psychiatric symptoms were compared pre- and post-intervention using the Positive and Negative Syndrome Scale, which assesses positive (7 items), negative (7 items), and general psychiatric symptoms (16 items) on a 7-point scale ranging from 1 to 7[9]. A reduction in total score reflects symptomatic improvement.

Nurse’s Observation Scale for Inpatient Evaluation scoring: The Nurse’s Observation Scale for Inpatient Evaluation (NOSIE) was administered by nursing personnel at baseline and after the 8-week intervention[10]. Symptom frequency was graded on a 5-point scale ranging from 0 to 4 (0 = none, 1 = occasional, 2 = moderate, 3 = frequent, 4 = persistent). The 30-item scale included 14 positive and 16 negative items. Positive factors (social competence, social interest, and personal hygiene) yielded a subtotal ranging from 0 to 112. Negative factors (agitation, psychotic symptoms, withdrawal, and depression) yielded a subtotal ranging from 0 to 128. The composite illness score was (128 + positive total - negative total), with a maximum score of 240. Higher total scores reflect better overall clinical status.

Stroop Color-Word Test performance: Stroop Color-Word Test (SCWT) outcomes (word/color/total correct responses) were measured pre- and post-8-week intervention[11]. Participants were instructed to read aloud words, colors, or color-words from a test card as quickly as possible within 45 seconds. The number of correct responses in each condition and the total score were recorded, with higher scores indicating better cognitive flexibility, selective attention, and the ability to suppress automatic response tendencies.

Information processing speed: To gauge processing speed, patients completed the Trail Making Test (TMT) and the Brief Assessment of Cognition in Schizophrenia (BACS-SC) before and after the 8-week intervention[12]. For the TMT, participants were instructed to connect numbers from 1 to 25 in ascending order as quickly and accurately as possible. Shorter completion times and fewer errors indicated better processing speed. For the BACS-SC, the number of correctly matched symbols to digits within a 90-second time limit was recorded, with higher scores reflecting faster information processing.

Assessment of social functional deficits: Patient social functioning was evaluated using the Inpatient Psychiatric Rehabilitation Outcome Scale (IPROS) at baseline and after the 8-week intervention. Each item was rated on a 5-point scale ranging from 0 (no symptoms) to 4 (most severe symptoms), with intermediate levels of 1 (mild), 2 (medium), and 3 (significant)[13]. The total score ranged from 0 to 144 and comprised five domains: Occupational therapy performance (maximum 32 points), daily living skills (maximum 32 points), interpersonal skills (maximum 24 points), hygiene practices (maximum 24 points), and engagement/motivation levels (maximum 32 points). Lower total scores indicated better social and functional outcomes.

Serum BDNF quantification: Baseline and post-treatment (8-week) venous blood draws (5 mL, fasting state) were performed. Samples were centrifuged, and the separated serum was subjected to BDNF level assessment using Human BDNF enzyme-linked immunosorbent assay test kits.

Statistical analysis

Statistical analyses were conducted using SPSS 20.0. Categorical variables were summarized as n (%), and continuous variables as mean ± Standard Deviation. Between-group differences were assessed for statistical significance using the χ² test for categorical variables and independent samples t-tests for continuous variables. Within-group changes before and after the intervention were analyzed using paired-samples t-tests. A significance threshold of 0.05 was applied.

RESULTS

Baseline characteristics

An analysis of the baseline data showed no significant differences between the control and intervention groups with respect to sex distribution, age, duration of illness, length of hospital stay, educational level, or types of antipsychotic medication used (all P > 0.05), indicating good baseline comparability (Table 1).

Table 1 Comparative analysis of baseline characteristics, n (%).

Categories	Control group (n = 59)	Research group (n = 68)	χ2/t value	P value
Sex			1.106	0.293
Male	35 (59.32)	34 (50.00)
Female	24 (40.68)	34 (50.00)
Age (years), mean ± SD	43.03 ± 8.72	40.03 ± 10.59	1.726	0.087
Disease course (years), mean ± SD	8.49 ± 3.91	7.94 ± 3.49	0.838	0.404
Hospital stay duration (months), mean ± SD	10.31 ± 2.82	10.56 ± 3.74	0.420	0.675
Educational background (years), mean ± SD	6.75 ± 2.64	7.09 ± 3.64	0.594	0.553
Main antipsychotics			2.986	0.225
Clozapine	42 (71.19)	40 (58.82)
Risperidone	6 (10.17)	14 (20.59)
Other	11 (18.64)	14 (20.59)

Comparative analysis of psychiatric symptom improvement

Changes in psychiatric symptoms assessed by the Positive and Negative Syndrome Scale are shown in Figure 2. At baseline, scores for positive, negative, and general psychiatric symptoms were comparable between the two groups (P > 0.05). Following treatment, the intervention group exhibited marked improvement, with significantly reduced scores across all symptom categories compared to controls (P < 0.05).

Figure 2 Comparison of symptom improvement profiles. ^aP < 0.05, ^bP < 0.01 (intra-group vs baseline); ^cP < 0.05 (inter-group comparison at matched time points). A: Positive symptom scores (PANSS) pre/post intervention; B: PANSS scores (negative symptoms) pre- and post-intervention across groups; C: Modifications in PANSS general psychiatric symptom scores before/after treatment.

Comparative analysis of NOSIE scores

Pretreatment comparisons indicated comparable NOSIE scoring patterns between the two groups (P > 0.05) (Figure 3). Following treatment, both groups showed increases in positive factor and total scores, with the intervention group demonstrating significantly higher post-treatment values than the control group (P < 0.05). Negative factor scores displayed progressive decline in both groups, reaching lower levels in the intervention group (P < 0.05).

Figure 3 Comparative analysis of Nurses’ Observation Scale for Inpatient Evaluation assessments. ^aP < 0.05, ^bP < 0.01 (intra-group vs baseline); ^cP < 0.05 (inter-group at equivalent timepoints). A: Changes in positive symptom scores (pre- vs post-intervention); B: Changes in negative symptom scores (pre- vs post-intervention); C: Total score variations across study periods.

Comparative analysis of SCWT performance

Table 2 compares SCWT outcomes between groups. At baseline, no significant between-group differences were observed in word, color, or total correct responses (P > 0.05). Post-intervention, the intervention group outperformed the control group in word and color tasks (P < 0.05). The control group showed isolated improvement only in word recognition (P < 0.05). Total correct response scores did not change significantly in either group after treatment, and no between-group differences were detected post-intervention (P > 0.05).

Table 2 Comparative assessment of Stroop Color-Word Test metrics, mean ± SD.

Categories	Control group (n = 59)	Research group (n = 68)	t value	P value
SCWT word recognition accuracy
Before intervention	65.53 ± 18.14	67.32 ± 19.12	0.539	0.591
After intervention	69.93 ± 13.00a	75.41 ± 15.17b	2.168	0.032
SCWT color recognition accuracy
Before intervention	41.49 ± 14.28	39.94 ± 13.53	0.628	0.532
After intervention	42.37 ± 13.70	49.94 ± 19.25b	2.517	0.013
Total correct responses
Before intervention	21.59 ± 10.91	21.50 ± 8.02	0.053	0.958
After intervention	23.80 ± 10.31	23.56 ± 9.67	0.135	0.893

^aP < 0.05.

^bP < 0.01 (intra-group vs baseline).

SCWT: Stroop Color-Word Test.

Cognitive processing speed assessment

Table 3 presents the information processing speed comparison. No baseline between-group differences were observed for TMT completion time or BACS-SC scores (P > 0.05). The control group showed no changes in performance after treatment (P > 0.05). Conversely, the intervention group demonstrated significantly shorter TMT completion times and higher BACS-SC accuracy compared with both baseline values and the post-treatment values in the control group (P < 0.05).

Table 3 Processing speed differences across groups, mean ± SD.

Categories	Control group (n = 59)	Research group (n = 68)	t value	P value
TMT completion time (seconds)
Before intervention	82.90 ± 29.66	82.41 ± 27.22	0.097	0.923
After intervention	83.88 ± 28.14	66.15 ± 23.34a	3.881	< 0.001
Accuracy count on BACS-SC
Before intervention	27.64 ± 11.35	27.76 ± 9.97	0.063	0.950
After intervention	30.22 ± 8.93	36.85 ± 12.23a	3.443	< 0.001

^aP < 0.01.

TMT: Trail Making Test; BACS-SC: Brief Assessment of Cognition in Schizophrenia.

Comparative assessment of social functional deficits

Table 4 presents the IPROS scale results for both groups. Baseline scores did not differ significantly between groups across any domain or in total (all P > 0.05). After the intervention, the control group showed significant deterioration in occupational therapy performance, daily living skills, and engagement/motivation, with higher total IPROS scores. In contrast, the intervention group demonstrated significant improvements across all domains, resulting in lower total scores and significantly better post-intervention outcomes compared with the control group (P < 0.05).

Table 4 Analysis of differences in social functional deficits, mean ± SD.

IPROS (points)	Control group (n = 59)	Research group (n = 68)	t value	P value
Occupational therapy performance
Pre-intervention	10.24 ± 2.90	10.34 ± 3.06	0.188	0.851
Post-intervention	8.12 ± 3.18a	6.54 ± 2.63b	3.064	0.003
Daily living skills
Pre-intervention	9.54 ± 2.73	8.66 ± 2.95	1.735	0.085
Post-intervention	6.95 ± 2.87a	5.41 ± 2.39b	3.299	0.001
Interpersonal skills
Pre-intervention	8.29 ± 3.29	8.15 ± 2.40	0.276	0.783
Post-intervention	7.51 ± 2.65	5.54 ± 2.39b	4.404	< 0.001
Hygiene practices
Pre-intervention	7.83 ± 2.81	7.51 ± 2.92	0.627	0.532
Post-intervention	6.41 ± 2.53	5.15 ± 2.14b	3.041	0.003
Engagement/motivation levels
Pre-intervention	10.19 ± 3.63	10.56 ± 3.72	0.565	0.573
Post-intervention	7.44 ± 2.90a	5.68 ± 2.51b	3.667	< 0.001
Total
Pre-intervention	46.10 ± 6.62	46.44 ± 5.94	0.305	0.761
Post-intervention	36.54 ± 6.33a	28.12 ± 6.10b	7.624	< 0.001

^aP < 0.05.

^bP < 0.01.

IPROS: Inpatient Psychiatric Rehabilitation Outcome Scale.

Serum BDNF concentrations

Serum BDNF levels are shown in Figure 4. No baseline differences existed between groups (P > 0.05). After intervention, BDNF levels increased significantly in both groups, with the intervention group exhibiting significantly higher post-treatment levels than the control group (P < 0.05).

Figure 4 Serum brain-derived neurotrophic factor concentration comparison. ^aP < 0.05, ^bP < 0.01 (intra-group vs baseline); ^cP < 0.05 (inter-group at equivalent timepoints). BDNF: Brain-derived neurotrophic factor.

Correlation of the change amplitude of BDNF with the reduction of negative symptoms and the shortening of TMT completion time

Table 5 presents the correlation between the change amplitude of BDNF and the reduction of negative symptoms or the shortening of TMT completion time, further evaluated based on the Pearson correlation coefficient. It was found that the BDNF change amplitude did not correlate significantly with the reduction of negative symptoms or the shortening of TMT completion time (P > 0.05).

Table 5 Correlation of the change amplitude of brain-derived neurotrophic factor with the reduction of negative symptoms and the shortening of Trail Making Test completion time.

Category	r	P value
BDNF change amplitude (points)	0.033	0.709
Reduction of negative symptoms (points)
BDNF change amplitude (points)	0.042	0.641
Shortening of TMT completion time (second)

BDNF: Brain-derived neurotrophic factor; TMT: Trail Making Test.

DISCUSSION

Schizophrenia has a lifetime prevalence of approximately 1% and is characterized by marked clinical heterogeneity[14]. Despite being the cornerstone of treatment, traditional antipsychotics show limited effectiveness against cognitive and negative symptoms[15]. The present study investigated whether adding exercise-based rehabilitation training to standard antipsychotic therapy and health education could enhance clinical outcomes in patients with schizophrenia.

Our findings indicate that exercise-based rehabilitation significantly reduced symptom severity across positive, negative, and general psychiatric domains. These results are consistent with the systematic review and meta-analysis by Kim et al[16], which demonstrated beneficial effects of exercise interventions on both negative and positive symptoms in schizophrenia. These clinical benefits, to a certain extent, are likely linked to exercise-induced hippocampal volumetric increases and serum BDNF upregulation[17]. Additionally, the sense of achievement and self-efficacy improvement brought about by exercise, as well as the naturally increased social interaction in group training, may subtly promote the remission of negative symptoms in patients with schizophrenia.

Furthermore, exercise-based rehabilitation training effectively alleviated symptoms in schizophrenia patients. NOSIE scores showed increased positive factors and total scores with concomitant reductions in negative factors. Cognitive performance improved as evidenced by higher SCWT accuracy in the word and color recognition tasks, along with faster TMT completion and higher BACS-SC correct responses. This might be attributed to the fact that the exercise-based rehabilitation training adopted in this study included route memorization walking and sandbag and table tennis drills, which are high-load cognitive-motor dual tasks, helping to specifically strengthen the executive control function of patients. Therefore, they are more effective for the SCWT sub-tasks that require complex cognitive control, while they perform less well in the relatively simple “total correct score” task. These findings align with the report by Nuechterlein et al[18], demonstrating exercise-related improvements in cognitive and occupational or educational functioning. The observed benefits may be mediated by exercise-induced hippocampal neuroplasticity, including enhanced neurogenesis and synaptogenesis. Expansion of hippocampal neural progenitor cell populations may further support gains in memory retention and learning capacity[19,20]. Moreover, while maintaining the mainstream aerobic and coordination training, the exercise-based rehabilitation training implemented in this study innovatively integrates social interaction, cognitive load, and dynamic task switching, forming a unique social-cognitive-motor integrated rehabilitation model. This rehabilitation model offers a replicable and highly participatory non-pharmacological intervention option for patients with schizophrenia.

IPROS assessments indicated that exercise-based rehabilitation training improved psychosocial functioning, including occupational therapy performance, daily living skills, interpersonal skills, hygiene practices, and engagement/motivation levels. These findings are consistent with those of Pérez-Romero et al[21], who reported significant improvements in executive functioning following exercise-based interventions in patients with schizophrenia. In addition, Dauwan et al[22] demonstrated that physical activity interventions improve clinical symptoms, psychosocial functioning, and affective states, further supporting the present results.

Furthermore, exercise-based rehabilitation training significantly increased serum BDNF concentrations in patients with schizophrenia, supporting the neuroplasticity hypothesis of exercise-based interventions. BDNF, a protein belonging to the neurotrophin family, is produced in the central nervous system and peripheral tissues. It promotes neuron survival, neurogenesis, and neural plasticity within the central nervous system. Studies suggest its involvement in memory (short-/Long-term) formation, executive control, cognitive processing, emotional stability, spatial orientation, and learning efficiency[23]. Physical activity increases BDNF both centrally and peripherally, as reflected in human serum/plasma studies[24]. Consistent with these mechanisms, Qigong exercise has been shown to improve cognitive dysfunction in stable schizophrenia[25]. Dunleavy et al[26] observed a correlation of moderate-vigorous exercise with decreased pro-inflammatory markers and ameliorated negative symptomatology in males with early-stage psychosis. Finally, the correlation assessment showed no significant association of the BDNF change amplitude with the reduction of negative symptoms in patients with schizophrenia or the shortening of TMT completion time. This may be because BDNF mainly promotes basic neural repair, while symptom improvement involves more complex brain network remodeling, resulting in a non-linear or indirect association between BDNF changes and specific behavioral outcomes.

Several limitations of this research should be acknowledged. First, as a single-center study with a small sample size, future research should expand to multi-center, large-sample analysis to improve the representativeness of the sample and increase the universality of the research results. Second, no causal relationship assessment has been established between the increase in BDNF levels and the functional improvement of patients with schizophrenia. In the future, longitudinal design and mediation analysis should be supplemented to clarify whether BDNF elevation serves as a causal indicator for the functional improvement of patients with schizophrenia.

CONCLUSION

In this study, exercise-based rehabilitation training was associated with improvements in psychiatric symptoms, cognitive performance, information processing speed, and social functioning. These benefits may be partially mediated by exercise-induced upregulation of serum BDNF.