The effects of dance intervention on clinical symptoms and cognitive deficits in hospitalized patients with chronic schizophrenia: a randomized controlled trial

Participants

Sixty hospitalized patients with chronic schizophrenia were recruited from the Fourth People’s Hospital in Chengdu, China. All participants maintained a stable medication regimen for at least 8 weeks before and during the dance intervention, with no changes to their antipsychotic medication type or dosage.

Inclusion criteria were: (1) met DSM-IV criteria for schizophrenia; (2) met the chronic criteria which possess 10–30 years of disease course (Harding, Zubin & Strauss, 1992); (3) aged between 30–60 years old; (4) no history of suicide attempts within the past 6 months (Andreasen et al., 2005).

Exclusion criteria were: (1) acute schizophrenia or in the onset period; (2) brain organic disorders (3) severe chronic underlying disease symptoms; (4) overweight and underweight; (5) significant cardiovascular, neuromuscular and endocrine or other somatic disorders; (6) space claustrophobia; (7) inability to comprehend the instructions in this study.

The study was registered in the Chinese Clinical Trials Register (ChiCTR2100049273) and approved by the local ethics committee of the University of Electronic Science and Technology of China (1061420210617005). All patients were duly informed of the study’s purpose, and written informed consents were obtained from all participants. The study was performed in accordance with the 1964 Declaration of Helsinki (World Medical Association, 2013).

Study design

Participants were randomly assigned to either the dance intervention group (DI) or the aerobic exercise group (AE). The study aimed to compare the effects of dance with those of aerobic exercise, focusing on the specific impact of the dance intervention. None of the participants had prior dance experience. Clinical symptoms and cognitive function were evaluated before and after the intervention. All evaluations were performed and interpreted by the same researchers or experienced clinicians, including psychiatrists or psychiatric inpatient nurses, who were blinded to the identity of the participants and the pre- and post-test results.

Both intervention programs, DI and AE, consisted of 50-minute sessions, three times a week, over the course of three months. The programs were designed with careful consideration of the participants’ abilities and safety. Previous research has shown that mild and moderate intensity physical activity can improve cognitive performance in older adults more effectively than vigorous activity (Wu et al., 2021). Given that chronic schizophrenia patients often experience severe extrapyramidal side effects from long-term medication, which impair physical flexibility and lead to disabilities. According to a previous review of the aerobic exercise for schizophrenia patients (Stanton & Happell, 2014), we finally set the intensity of the intervention at 40%–50% of the target heart rate.

To ensure comparable exercise levels across both groups, participants were required to wear exercise bracelets to record their heart rate. The volume of exercise was measured using volume of oxygen (VO₂) max, with the Karvonen formula used to calculate each participant’s target heart rate (HR). The formula, which is based on the maximum heart rate (220 minus age) (Goldberg L & Kuehl, 1988), helped ensure that all participants were exercising at a moderate intensity. The formula is as follows. (1)${\displaystyle \text{Target Heart Rate}=\left[\left(max\text{HR-resting HR}\right)\times \left(40\text{\%}\text{to}50\text{\%}\right)\text{Intensity}\right]+\text{resting HR}.}$

The goal was for patients in the dance intervention group to achieve a target heart rate corresponding to an exercise intensity level. The target HR range for the dance group was calculated between 110 and 140 beats per minute, based on their pre-intervention HR, which was recorded and averaged. Changes in heart rate for both groups, along with the target rates, are detailed in the Supplementary File SA.

Measures

Clinical assessment

The Positive and Negative Syndrome Scale (PANSS), is an internationally recognized authoritative tool for measuring clinical symptoms in patients with schizophrenia, with high reliability and validity ranging from 0.83–0.99 (Kay, Fiszbein & Opler, 1987). The Positive scale (PANSS-P) mainly measures the state of schizophrenic patients in delusions, hallucinations, and some other chaotic thinking manifestations. In contrast, the negative scale (PANSS-N) mainly measure cognitive, emotional, and social deficits. The General Psychopathology Scale (PANSS-G) measures general clinical symptoms such as anxiety and tension. The subscale scores ranges from 1 to 7, with higher scores indicating more severe symptoms. The total score (PANSS-total) was the sum of the three sub-scale scores, ranging from 30–210.

Table 2:

Difficulty scores for the dance clips used in the intervention.

Clip	Time				Space			Strength	Total score	Difficulty degrees
	Duration	Rate of motion	Movements in one beat	Motion switch times	Orientation difficulty score	Orientation switches	Orientation difficulty switches	Body parts involved
1	0.25	0.36	0.245	0.075	0.25	0.06	0.075	0.556	1.871	low
2	0.50	0.27	0.153	0.275	0.417	0.09	0.150	0.667	2.522	medium
3	0.34	0.51	0.51	0.175	0.458	0.16	0.175	0.667	2.995	high

DOI: 10.7717/peerj.19840/table-2

Clinical behavior change was assessed by the Nurses’ Observation Scale (NOISE) (Lyall, Hawley & Scott, 2004), which evaluated symptoms such as social ability, interest, personal hygiene, irritability, psychotic performance, retardation, and depression exhibited by patients over the last two weeks. All assessments were performed by experienced psychiatrists or psychiatric nurses who were blinded to the group allocation.

Data analysis

Data were analyzed using SPSS version 26.0 (IBMCorp., Armonk, NY, USA). Group differences in demographic variables were assessed using independent-samples t-tests and chi-square tests, including gender, age, years of education, height, weight, BMI, illness duration, age at onset, medication dosage, and number of intervention sessions. Given the small sample size, non-parametric tests were also employed to validate the results. The Mann–Whitney U test (Ruxton, 2006) was used to assess baseline differences between the dance intervention and aerobic exercise groups.

Outliers beyond three standard deviations were excluded, after which a 2 ×2 repeated-measures ANOVA was conducted to examine the clinical symptoms, cognitive functions, and physiological indicators between the two groups. To complement the repeated-measures analysis, we used generalized estimating equations (GEE) for non-parametric analysis, which are particularly useful in small-sample longitudinal studies (Hanley, 2003; Gosho et al., 2023; Tada & Sato, 2024). A Bonferroni correction was applied for multiple comparisons (Simes, 1986) to ensure the robustness of our findings.

Finally, we used partial correlation analysis to examine the relationships between clinical symptoms, cognitive function, and physiological indicators, adjusting for potential confounders including age, gender, education level, height, and weight.

Group comparison at baseline

Between September and December 2021, a total of twenty-eight participants dropped out during the intervention (Fig. 3). Consequently, thirty-two patients completed the full intervention program, including baseline and 3-month assessments of clinical symptoms and cognitive function. These 32 participants were included in the final analysis. There were no significant differences between dance intervention and aerobic exercise groups in terms of age, gender, education years, duration of illness, and medication dosage in chlorpromazine (CPZ) (Table 3). Furthermore, no significant differences were found in clinical symptoms or cognitive functions at baseline between dance group and aerobic exercise group (Table 4). on-parametric tests, consistent with the independent-samples t-tests, also revealed no significant baseline differences in cognitive function between the two groups (Supplementary File SC).

Clinical symptom

As shown in Table 5, significant main effect of time main were observed for PANSS-total (p < 0.01), PANSS-N (p < 0.001), and NOISE scores (p < 0.001).

After 3 months of intervention, both the dance and aerobic exercise groups showed significant reductions in PANSS total, PANSS-N, and NOISE scores, indicating improvements in the clinical symptoms of schizophrenia.

Cognitive function

Repeated-measures ANOVA revealed significant main effect of time for TMT, CF, SS, DS, BVMT-R, and HVLT-R (Table 6). Compared with baseline, these measurements increased significantly in both groups after 3 months of intervention indicating that both dance and aerobic exercise interventions improved cognitive function in schizophrenia. Significant Time × Group interaction effects were observed in CPT-IP and HVLT-R, with a notably improvement in the dance intervention compared to aerobic exercise group (Fig. 4). In addition, the results from the GEE also showed significant interaction effect in CPT-IP and HVLT-R, further supporting the reliability of these findings (Supplemental File SD).

Physiological indicators

There were significant Time × Group interaction effects in mean corpuscular hemoglobin (MCH, p = 0.027), mean corpuscular hemoglobin concentration (MCHC, p = 0.048), plateletcrit (PCT, p = 0.037), cystatin C (CYS-C, p = 0.008), total bilirubin (TBIL, p = 0.041) (Table 7).

As shown in Fig. 5, CYSC levels significantly increased in dance intervention group. MCH and MCHC significantly decreased in the aerobic exercise group, while no significant changes were observed in the dance exercise group. PCT levels significantly increased in the aerobic exercise group has a significant increase. Additional results are provided in Supplemental File SE and SF.

Correlation between clinical symptom, cognitive function and physiological indicators

As shown in Fig. 6, a significant positive correlation was found between changes in TBIL and PANSS-P in both the aerobic exercise group and the combined group. Similarly, a significant positive association between changes in CPT-IP and CYS-C was observed in the dance intervention group and in the total sample.

Both interventions induced beneficial effects on clinical symptoms and cognitive functions

Both interventions resulted in significant improvements in negative symptoms, information processing speed, working memory, short-term memory, and visual memory in schizophrenia. These improvements may be attributed to the shared elements of physical activity in both dance and aerobic exercise.

Dance and aerobic exercise involve whole-body movements that can enhance brain plasticity and offer potential therapeutic for individuals with schizophrenia. Previous studies have shown that interventions significantly increased hippocampal volume, a brain area associated with memory functions in older adults (Pajonk et al (2010)) and chronic schizophrenia patients (Erickson et al., 2011). Furthermore, both interventions are known to elevate levels of brain-derived neurotrophic factor (BDNF), a protein critical for neuroplasticity and brain adaptation (Numakawa, Odaka & Adachi, 2017; Nieto, Kukuljan & Silva, 2013; Lee et al., 2001). Reduced levels of BDNF are frequently reported in chronic schizophrenia patients and are associated with impairments in negative symptoms and verbal working memory (Toyooka et al., 2002; Niitsu et al., 2011). Both dance intervention and aerobic exercise were found to increase serum BDNF levels (Kim et al., 2018; Griffin et al., 2011), suggesting that the elevation of BDNF may represent shared underlying mechanism. In addition, both interventions require sustained musculoskeletal exertion and provides strength to cardiorespiratory system (Schmitt et al., 2018). Several studies have demonstrated that both the two intervention enhanced cardiorespiratory fitness in schizophrenia patients (Cheng, Sun & Yeh, 2017; Kimhy et al., 2014; Scheewe et al., 2013), which has been linked to improvement in positive (Wang et al., 2018), negative (Bryl et al., 2022), and general symptoms (Gökcen et al., 2020) as well as cognitive functions (Chen et al., 2016; Firth et al., 2017; Maurus et al., 2023).

Dance intervention induced specific cognitive improvement

Only the dance intervention led to significant improvements in attention and verbal memory. This suggests that dance may uniquely enhance the integration of sensory and motor information. Sensorimotor deficits are considered to be a core feature of schizophrenia and a key contributor to cognitive dysfunction (Javitt & Freedman, 2015; Kaufmann et al., 2015). Unlike general aerobic exercise, dance requires higher levels of sensorimotor integration and attention load. Performing dance movement requires sustained attention to auditory cues and instant synchronization of auditory, visual, and motor sensory information (Năstase, 2012). Previous studies have shown that dance improves attention switching and cognitive performance (Kattenstroth et al., 2010; Coubard et al., 2011), indicating that dance intervention may strengthen sensory-motor function of patients, and further enhance their attention control.

Additionally, the observed improvement in verbal memory may be linked to the semantic processing inherent in dance. Dance involves both action semantic (Van Elk, Van Schie & Bekkering, 2014) and music engagement (Gustavson et al., 2021), both of which rely on multisensory integration. A dance phrase typically consists of a coherent sequence of movements with clear beginnings and endings, forming a narrative structure (Grosbras, Beaton & Eickhoff, 2012). Comprehending this sequence draws on neural systems involved in processing meaning, reinforcing the connection between movement comprehension and language understanding (Bachrach, Jola & Pallier, 2016). Moreover, the musical component of dance also plays a linguistic role. Music engagement has been shown to support verbal memory through mechanisms similar to language processing (Jakobson et al., 2008; Ho, Cheung & Chan, 2003; Chan, Ho & Cheung, 1998; Jakobson, Cuddy & Kilgour, 2003; Brandler & Rammsayer, 2003). Verbal memory plays a critical role in the comprehension of linguistic texts and understanding the intentions behind spoken words in everyday life (Yıldız, Oğuzhanoğlu & Topak, 2023), which may explain the dance intervention’s positive effect on verbal memory in schizophrenia patients.

Taken together, the complexity and multimodal nature of dance may provide a unique platform for integrating primary sensorimotor information, particularly in patients with schizophrenia, thereby enhancing the brain’s capacity to process higher cognitive functions.

The two interventions have discrepant effects on the physiology level

In the dance intervention group, CYS-C levels significantly increased and were positively correlated with improved attention, as measured by CPT-IP. CYS-C is an endogenous cysteine proteinase inhibitor found abundantly in cerebrospinal fluid (Mussap & Plebani, 2004). It has been associated with neuroprotection and cognitive outcomes (Cao et al., 2022) in disorders such as Alzheimer’s disease (Sundelöf et al., 2008). Lower levels of CYS-C have also been observed in schizophrenia (Mulugeta, Suppiah & Hyppönen, 2023; Owen, Sawa & Mortensen, 2016). The post-intervention increase in CYS-C may reflect improved neural protection and cognitive function in patients receiving dance therapy.

In the aerobic exercise group, a decrease in total bilirubin (TBIL) was observed. Bilirubin, an antioxidant, is implicated in the pathophysiology of schizophrenia (Radhakrishnan et al., 2011; Yao, Reddy & Nan Kammen, 2000; Widschwendter et al., 2016; Hankøet al., 2005). Although low bilirubin levels can reduce oxidative stress, excessive reductions may diminish antioxidant capacity (Wang et al., 2023). We observed a positive correlation between TBIL levels and positive symptoms, suggesting that lower TBIL was associated with symptom reduction, possibly through improved oxidative balance. Additionally, MCH and MCHC levels decreased in the aerobic group, while PCT increased. MCH is linked to treatment-resistant schizophrenia (Cheng et al., 2024), and MCHC is associated with red blood cell aggregation and metabolic syndrome (Barshtein et al., 2004; Toker et al., 2005). Elevated PCT levels reflect increased platelet activity and inflammation (Yu et al., 2020). These findings suggest that aerobic exercise influences peripheral physiological systems, potentially modulating inflammatory or metabolic pathways relevant to schizophrenia.

In summary, the divergent patterns of physiological change between groups may reflect different underlying biochemical mechanisms triggered by dance and aerobic exercise.

Limitation

This study has several limitations. First, a relatively high attrition rate reduced the final sample size. Future studies should aim to recruit larger samples and adopt strategies to minimize dropout. Second, although the aerobic exercise group served as an active control, the absence of a non-intervention (blank) control group limits the ability to isolate the specific effects of dance. Third, fluctuations in physiological markers may have been influenced by seasonal changes or dietary variations, which were not controlled for. Additionally, premorbid variables such as substance use (Prat et al., 2021) and history of suicide attempts (Le et al., 2024), both of which can influence cognition, were not accounted for. Moreover, given the distinct influence of circadian rhythms on individuals with schizophrenia (Adan et al., 2024), future studies should adopt more rigorous controls over the timing of interventions. Finally, the short duration of the intervention precludes conclusions about long-term effects, which future longitudinal studies should address.

Outlook

This study supports the potential of dance as a practical, cost-effective adjunctive therapy for schizophrenia. We developed a structured, feasible dance program designed for implementation by healthcare providers, such as psychiatrists and psychiatric nurses, without requiring professional dancers. Its simplicity and standardization enhance its clinical utility.

Nevertheless, individualized approaches are essential. In some patients, high exercise intensity may exacerbate extrapyramidal symptoms. Tailoring intervention intensity to the individual’s physical and psychiatric condition is critical to maximize the benefit.

Future research should emphasize interdisciplinary collaboration, integrating neuroscience, psychology, and rehabilitation medicine. Special attention should be given to identifying the neural mechanisms by which dance influences cognition—particularly the modulation of sensorimotor integration in schizophrenia.