1.1 Cognitive disorder in type 2 diabetes: mechanisms and manifestations

Type 2 diabetes mellitus (T2DM), a chronic metabolic disorder, is increasingly recognized for its association with cognitive decline, ranging from mild cognitive impairment (MCI) to dementia, beyond its established risks for cardiovascular, renal and ophthalmic complications.¹,² Individuals with T2DM are 1.5–2 times more likely to develop cognitive dysfunction compared to non-diabetics.³,⁴ The underlying mechanisms are multifactorial, involving insulin resistance, chronic hyperglycemia, oxidative stress, cerebrovascular pathology and inflammation.⁵,⁶ Psychological factors such as depression, anxiety and poor self-care further contributes to cognitive burden.⁷ The hormonal imbalances particularly in the insulin and insulin like growth factor (IGF) axis.⁸ Neuroinflammation and oxidative stress disrupts brain regions responsible for memory, executive function and attention.⁹,¹⁰ Cognitive deficiencies have been further exacerbated by insulin pathways, effects the synaptic plasticity and neural communication.¹¹ These changes develop gradually, with midlife metabolic dysfunction often leading to late-life cognitive impairment.⁴,¹²

Cognitive impairment in T2DM commonly effect working memory, executive function, attention and processing speed,²,¹³ impacting self-management behaviours such as insulin dosing, diet regulation and appointment adherence.¹⁴ These deficits stem not only from aging but also from diabetes related metabolic abnormalities. Neuroimaging reveals cortical thining, hippocampal atrophy and impaired connectivity in the prefrontal and parietal regions of T2DM individuals,¹⁵,¹⁶ often accompanied by slowed gait, poor balance and delayed reaction, increasing fall risk and loss of independence.¹,¹⁷ Psychological factors like social withdrawal, anxiety and depression further impair cognition, creating a feedback loop that undermines cognitive control and self-care.⁶,⁷

1.2 Recent advances in rehabilitation approaches and role of CMDT

Cardiovascular exercise, resistance training, and improving balance are the main goals of traditional physiotherapy for individuals with type 2 diabetes. These techniques help maintain cardiovascular health and glycemic management, but they frequently fall short in addressing cognitive deficiencies.¹⁸,¹⁹ As an alternative, cognitive-motor dual task (CMDT) training is a potential approach that concurrently targets higher order cognitive functions and motor coordination. Compared to single-task training, CMDT engages frontal executive brain networks more strongly by requiring participants to perform a physical and cognitive task simultaneously, such as walking while reciting alternative numbers or performing basic arithmetic problems.²⁰ Research has demonstrated that this paradigm improves mental flexibility, split attention, and dual-task gait performance in diabetic patients by simulating real-world functional demands.²¹ There is recent evidence that CMDT can result in significant improvements in neuropsychological profiles, especially for those who have early cognitive abnormalities. According to Şekerci et al.,²⁰ fuzzy clustering analysis has identified unique cognitive abnormalities in T2DM, each of which responds differently to CMDT therapies. This suggests that tailored approaches are necessary.

1.3 Vision feedback restriction as sensory modulation strategy

Vision feedback restriction is incorporated into CMDT as a new improvement. In order to induce reliance on alternate sensory modalities including proprioception and vestibular input, this strategy involves reducing visual input during training activities, either by partial occlusion or closed eye workouts.⁹,¹⁵ For individuals who frequently exhibit deficiencies in visual spatial processing and balance control as a result of diabetic neuropathy or retinopathy, limiting vision under supervision encourages sensory reweighting.²² By increasing attentional demands and activating underused brain pathways, vision-restricted CMDT is hypothesized to support neuroplastic adaptation; however, direct evidence remains limitied in T2DM population. Research using functional connectivity analysis has shown that this type of training can enhance coordination between sensorimotor and cognitive networks, particularly in those with cognitive impairment associated with T2DM.¹⁷,²³ Additionally, the posterior parietal cortex and cerebellum, which are involved in spatial navigation, motor timing, and mistake correction, may become more active when visual signals are removed. In people with diabetes, proprioceptive acuity is frequently impaired; this improves it.⁸,¹⁶ CMDT promotes neuroplasticity by simultaneously engaging motor and cognitive domain, thereby enhancing brain connectivity and function. Fang et al.,¹⁵ demonstrated that cognitive dysfunction in T2DM is unidirectional but can be modulated through interventions that alters the brain connectivity, supporting the potential of Dual-task training to induce beneficial neuroplastic changes.

1.4 Addressing the clinical challenges in T2DM cognitive disorder

Cognitive dysfunction in T2DM is rarely isolated. It frequently coexists with psychological anguish, sensory deficits, and motor instability. These interrelated deficiencies put a greater strain on caregivers and healthcare systems and provide serious obstacles to efficient illness management.¹⁴,⁶ Rehabilitation techniques must therefore focus on enhancing functional integration across the cognitive, physical, and emotional domains rather than only treating specific symptoms. In this sense, CMDT with vision feedback restriction shows promise. This technique may provide greater all encompassing advantages than traditional therapy since it challenges executive control, sensory coordination, and physical execution all at once. It is especially appropriate for people who have both postural instability and cognitive impairment, which is a prevalent symptom in older diabetic populations.¹,¹³ Moreover, integrating cognitive and motor training in settings that replicate everyday tasks may improve skill transferability and adherence. Maintaining long-term inde-pendence and lowering secondary problems such medication errors, falls, and readmissions to the hospital are critical outcomes of this.²⁴,⁷

1.5 Research gapand need for randomized controlled trial

Cognitive impairment is a frequent and progressive complication in individual with T2DM, linked to both structural and functional brain changes.¹,⁸,⁹ Despite this, intervention studies remain limited. While cognition and motor training have shown promise independently, dual-task training particularly CMDT offers a more ecologically valid effective approach.¹⁵,²³ The cognitive deficits have been consistantly reported, there are limited effective and scalable interventions to mitigate cognitive decline.⁵,⁶,¹³,¹⁵,²³ Current literature emphasizes the prevalence and risk factors of cognitive decline in T2DM.²,⁶,¹⁴ No RCTs have examined the added benefit of visually restricted CMDT, highlighting a clear gap. An RCT is essential to extablish causality and assess its efficacy.¹⁰,¹⁷,²²

1.6 Study rationale for vision feedback restricted cognitive-motor dual task traning

Innovative, integrative rehabilitation techniques are of vital importance especially in light of the known link between type 2 diabetes and cognitive decline and the shortcomings of existing therapy methods. In order to bridge this gap, vision feedback restricted CMDT combines multimodal altering with dual-task cognitive stimulation to promote neuroplasticity and functional recovery. Thus, the purpose of this study is to assess the effectiveness of this exploratory integrative approach in people with MCI and T2DM. Based on behavioural, neurobiological, and functional data, the method aims to guide treatment procedures that close the gap between managing metabolic diseases and cognitive rehabilitation.

1.7 Study objectives

Primary Objective to ascertain whether persons with type 2 diabetes and mild cognitive impairment can improve their overall cognitive performance through vision feedback-restricted cognitive-motor dual task training.

Secondary Objective to evaluate gains in motor function, task execution accuracy, and executive function in comparison to traditional physiotherapy therapies.

2.1 Patients and public involvement

Participants were not involved in the planning, recruiting, or execution of the study. To improve workout delivery, session duration, and the acceptance of visual feedback restriction, a small sample of eligible participants provided comments during the pilot phase. Participants helped to improve the intervention even though they were not directly involved in the main study process. To make sure they were aware of the findings and their part in influencing the study's methodology, each participant received a tailored feedback report at the end of the study that summarized the findings.

2.2 Ethical approval and clinical trial registration

The study received ethical approval from the institutional ethics committee of the Apollo Institute of Medical Sciences and Research in Chittoor, Andhra Pradesh, India (permit date: December 30, 2023; ethics committee number: PG/35/IEC/AIMSR/2023).The clinical trial was registered with the Indian Clinical Trials Registry (CTRI/2024/01/061956). Each participant gave written informed consent before to participation in compliance with national research ethical guidelines and the Declaration of Helsinki.

2.3 Study design

This single-blind randomized controlled experiment (RCT) was carried out to evaluate the impact of cognitive motor dual-task training (CMDT) on the cognitive performance of individuals who have type 2 diabetes mellitus (T2DM).

2.4 Trial design

A two-arm, parallel-group, single-blind randomized controlled trial (RCT) with a 1:1 allocation ratio was carried out. In order to determine whether the intervention group getting vision feedback restricted cognitive-motor dual task training (VFR-CMDT) showed more improvements in cognitive outcomes than a control group receiving traditional Cognitive-motor physiotherapy, the experiment used a superiority paradigm.

2.5 Changes to the trial protocol

The protocol was not significantly altered after it was started. The IRB-approved protocol included pre-specified time points, analytic approaches, and primary and secondary outcomes. For transparency, any deviations or administrative adjustments (such extending the recruiting period) were recorded; nonetheless, this had no bearing on the integrity of the results.

2.6 Trial setting

The study was carried out at the Apollo District Headquarters Hospital's Physiotherapy Outpatient Depart-ment in Murukambattu, Chittoor-517001, Andhra Pradesh, India. The study was carried out between July 2022 and February 2023.The Apollo District Headquarters Hospital's Physiotherapy Outpatient Department in Murukambattu has dual-task training modules, balancing platforms, and cognitive testing stations.

2.7 Eligibility criteria

Inclusion Criteria: Individuals who were between 40 and 70 years old, had been diagnosed with Type 2 Diabetes Mellitus for at least 5 years with HbA1c level more than 7 mmol/L, and had mild cognitive impairment (defined as a Montreal Cognitive Assessment (MoCA) score of less than 26) were eligible to participate. Other requirements included having a stable glycemic profile for at least four weeks before trial admission, being mobile, and being able to follow spoken instructions.

Exclusion criteria: Individuals with dementia or serious neurodegenerative illness, a history of stroke, brain damage, or major psychiatric problem, and significant uncorrected visual, vestibular, or hearing impairments were among the exclusion criteria. Moreover, those who had engaged in physical or cognitive therapy in the previous six months.

2.8 Site and personnel criteria

The investigation was carried out at the Physiotherapy Outpatient Department of the Apollo District Head-quarters Hospital in Murukambattu. The licensed physiotherapists who conducted the interventions had at least five years of clinical experience and had received specific training in dual-task cognitive-motor therapy. Adverse effects (such as lightheadedness and mental exhaustion) were tracked using a structured adverse event form, and all sessions were monitored. There were no significant negative effects or withdrawals.

2.9 Intervention

2.10 Outcome measures

2.11 Sample size calculation

G∗Power 3.1 was used to calculate the sample size. A minimum of 52 participants was required to achieve 80 % power at α = 0.05 for a one-tailed paired-samples t-test (dz = 0.35), and similar parameters indicated a requirement of approximately 64 participants (32 per group) for a two-tailed independent samples t-test (d = 0.5). To account for the potential attrition a large sample was recruited. A total of 105 individuals wereinitiallyscreened for eligibility. Follwing screening, 94 participants met the inclusion criteria and were subsequently randomized into two equal groups, Groups A (VFR-CMDT - n = 47) and Group B (CV-CMT - n = 47).During intervention period, a total of 14 participants (7 from each group) withdrew, resulting overall attrition rate of approximately 15 %. Consequently, 80 participants (40 per group) complete the study and were included in final analysis. CONSORT Flow diagram Fig. 1 shows the enrollment, randomization, follow up and analysis of the participants of the Group A and B.

Close

Fig. 1 CONSORT FLOW diagram showing participant recruitment, randomization and study completion.

Export to PPT

2.12 Randomization

Generation of Sequences An impartial statistician used permuted block randomization (block size = 4) to construct a random sequence on a computer.

Concealment of Allocation A third party who was not engaged in the recruitment or assessment process placed random assignments in sequentially numbered, opaque, sealed envelopes (SNOSE).

Implementation Following baseline assessment, participants were enrolled and assigned to interventions by an independent research coordinator. The allocation sequence was not disclosed to the intervention team.

The Blinding The hypothesis and the comparative nature of the interventions were concealed from the participants. Due to the nature of physical rehabilitation, physiotherapist and the participants could not be blinded, introducing the risk of performance and detection bias. However, to mitigate this, outme assessors and statisticians were blinded and participantsreceived non-comparative descriptions of interventions.

2.13 Statistical analysis

The IBM SPSS Statistics version 28 was used to examine the data. Per-protocal analysis was performed to evaluate the effectiveness of an intervention under ideal condition by including only the participants who fully adhered to the study protocol. Shapiro-Wilk and Levene's tests were used to determine whether the variances were normal and homogeneous, respectively. Paired and independent t-tests were used for within- and between-group comparisons, and repeated measures ANOVA with time and group as factors was used to analyze the primary outcomes. The threshold for significance was = 0.05. Analyses of post hoc power and effect sizes (Cohen's d, partial η²) were performed to assess the statistical and clinical significance.

3.1 Data preparation and descriptive statistics

Group A (n = 40) and Group B (n = 40) did not differ significantly on any of the baseline parameters, such as gender distribution, age, duration of diabetes, education, HbA1c, BMI, and cognitive tests (MoCA, TMT- B, SCCs) (all p > 0.05). The two groups’ gender distributions were similar. There were 25 males (62.5 %) and 15 females (37.5 %) in Group A, and 26 males (65 %) and 14 females (35 %), in Group B. There was no discernible gender disparity across the groups, as indicated by the statistically equivalent proportions. Baseline demographic and clinical characteristics are presented in Table 1. Fig. 2 shows the p-values in baseline demographic and clinical Characteristic.

Close

Fig. 2 Baseline demographic and clinical characteristics.

Export to PPT

3.2 Normality test and homogeneity of variance

The normality of the outcome variables (MoCA, TMT-B, and SCCs) at all time points was evaluated using Shapiro-Wilk tests and Quantile-Quantile plot (Q-Q plots), a graphical representation is given in Fig. 3 to assessed that data set follows a particular theoritical distribution. The findings supported parametric testing since there was no discernible departure from normalcy across groups or time points (p > 0.05 for all). All metrics at all time points showed homogeneity of variances between groups, according to Levene's tests (p > 0.05). The Shapiro-Wilk and Levene's test p-values, which show the normality and homogeneity of variance across groups and time points, are shown in Table 2.

Close

Fig. 3 Q–Q Plots for Normality Assessment with Shapiro-Wilk and Levene's Test p-values.

Export to PPT

3.3 Primary efficacy analysis of MoCA, TMT-B and SSCs scores

Global cognitive function (MoCA) showed significant improvements in both groups from baseline to the fourth week, with additional modest benefits continuing into the eighth week. At both post-intervention time points, Group A consistently scored higher than Group B. However, the differences were small, indicating that both therapies were successful and that Group A had a better cognitive advantage. Both groups showed improvement in terms of executive function (TMT-B) as completion times reduced significantly; however, Group A showed faster and higher increases, especially by the eighth week, when the performance gap increased, indicating a more apparent benefit for Group A. Subjective cognitive complaints (SCCs) also decreased significantly over time in all groups, but by the eighth week, Group A's scores had decreased more than Group B's, suggesting both objective improvements and a larger perceived cognitive gain from the intervention, as shown in Table 3. Normality Assessment (MoCA, TMT-B, SCCs) display group-wise means with error bars (±SD) Corresponding p-value from independent t-tests are indicated in Fig. 4.

Close

Fig. 4 Cognitive Outcomes at baseline, post intervention and follow-up for Group A and Group B.

Export to PPT

3.4 Mixed design ANOVA result for cognitive and subjective outcome

A mixed-design ANOVA was used to compare groups A and B and the impact of time on cognitive results (Baseline, Week 4, Week 8). With no group or interaction effects (p > 0.24), time had a high main effect on MoCA scores (F(2,76) = 125.3, p < 0.001), demonstrating significant progress over time in both groups. Similarly, there were no group differences or interactions, and time significantly improved executive function (TMT-B; F(2,76) = 180.5, p < 0.001). Over time, subjective cognitive complaints (SCCs) similarly declined (F(2,76) = 205.7, p < 0.001), with Group A showing a non-significant tendency toward faster improvement (p = 0.09). These findings reinforce that although there were time based improvements within each group, the absence of statistically significant group by time interaction effects (all p > 0.24) limits conclusion about the comparative superiority of VFR-CMDT. as shown in Table 4. Despite showing a numerical difference in favor of Group A, a p-value of 0.09 falls short of the standard cutoff point for statistical significance (p < 0.05).

3.5 ANOVA within and between groups comparisons over time

Table 5 displays descriptive data (Mean ± SD) for every result at every time point. Both Group A and Group B showed statistically significant gains in all measures (MoCA, TMT-B, and SCCs) from Baseline to Week 4 and from Week 4 to Week 8 (all p < 0.001), according to within-group analyses. Both groups had comparable patterns of cognitive gain over time, though, since between-group comparisons at each time point failed to produce statistically significant differences (all p > 0.05).

3.6 Post-hoc power analysis - group comparison over time

All variables demonstrated minimal between-group disparities at baseline, indicating well-matched cohorts. The initial effect size for MoCA was small (Cohen's d ≈ 0.25), progressively increasing to (d ≈ 0.50) by the eighth week, suggesting of delayed cognitive enhancement. Post-intervention, the effect size decline to a negligible level (d < 0.20). Executive function, assessed via TMT-B, exhibited continuous improvement, with effect sizes rising from (d = 0.17) at baseline to approximately (d = 0.31) at week four and (d ≈ 0.53) at week eight. Subjective cognitive complaints showed minimal variation at baseline (d = 0.15), and post-intervention data were available for analysis. While sufficient to detect moderate within-group effects, the study may have been underpowered to identify small but potentially meaningful between-group differences. Future trial should consider the power calculation based on observed effect szes (d = 0.3–0.6) for group comparision. Overall,as shown in Table 6 the intervention was favored by objective measures that demonstrated growing between-group effects over time.

Finally all outcome measures showed a significant increase in cognitive function over time for both groups, refer Fig. 5. Although Group A continuously performed better than Group B, the differences between the two groups were not statistically significant and were only slight. With Group A displaying a numerically greater improvement, these results imply that both therapies were successful.

Close

Fig. 5 Cognitive function improved at baseline, post intervention and follow-up for Group A and Group B.

Export to PPT