Recent Advances in Stroke Rehabilitation & Physiotherapy — Evidence-Based Research (2024–2026)

🏆 TIER 1 — Cochrane Systematic Reviews & Meta-Analyses (Highest Evidence)

1. Physical Rehabilitation for Function & Mobility After Stroke (Cochrane, 2025)

Todhunter-Brown A et al. Cochrane Database Syst Rev. 2025. DOI: 10.1002/14651858.CD001920.pub4

• Physical rehabilitation significantly improves activities of daily living (ADL) and motor function vs. no rehabilitation — with evidence of long-term benefits (low-certainty evidence due to study heterogeneity).
• Gait velocity improvement showed moderate-certainty evidence (SMD 0.23).
• No single approach (e.g., Bobath, task-specific, motor learning) was conclusively superior to another — what matters most is dose and delivery of therapy.
• Massive update: 267 studies across 36 countries. Half conducted in China.

Clinical implication: Intensity and task-specificity of therapy matter more than which named approach is used.

2. Virtual Reality for Stroke Rehabilitation (Cochrane, 2025)

Laver KE et al. Cochrane Database Syst Rev. 2025. DOI: 10.1002/14651858.CD008349.pub5

• VR improves upper limb function (SMD 0.20, low-certainty), balance (SMD 0.26, low-certainty), and reduces activity limitation (SMD 0.21, moderate-certainty).
• Adding VR to usual care provides additional benefit.
• Little to no effect on participation and quality of life.
• Wide range of applications tested: immersive VR glasses to non-immersive gaming platforms.

Clinical implication: VR is a viable adjunct for upper limb and balance training; moderate evidence for reducing activity limitation.

3. Electromechanical/Robot-Assisted Gait Training After Stroke (Cochrane, 2025)

Mehrholz J et al. Cochrane Database Syst Rev. 2025. DOI: 10.1002/14651858.CD006185.pub6

• Combined electromechanical + physiotherapy probably increases odds of independent walking (OR 1.65; moderate-certainty evidence).
• Modest increase in walking velocity (+0.05 m/s; moderate-certainty).
• No increase in walking capacity (6-minute walk test; high-certainty evidence — important null finding).
• Safe — no increased risk of dropout or death.

Clinical implication: Exoskeleton/robotic gait trainers improve independence in walking but do not substantially increase walking speed or endurance — combine with conventional physiotherapy.

4. VR + Exercise for Balance & Walking in Chronic Stroke (2024)

Krohn M et al. J Med Internet Res. 2024. DOI: 10.2196/59136

• VR training shows large effect on balance (SMD 0.51) and moderate effect on walking (SMD 0.31) in chronic-phase stroke survivors.
• Level of immersion (VR glasses vs. screen) did not significantly influence outcomes.
• Applicable to ambulatory working-age stroke rehabilitees.

Clinical implication: VR exercise is effective even in the chronic phase — important for community-based rehab programs.

5. Physical Therapy Modalities for Severe Stroke (Systematic Review, 2024)

Roesner K et al. Syst Rev. 2024. DOI: 10.1186/s13643-024-02676-0

• Evidence for PT in severe stroke is conflicting and uncertain (moderate to low quality).
• Covers 20 PT interventions including positioning, mobilization, and splinting.
• Major gap: therapeutic modalities (intensity, frequency, duration) are poorly reported.

Clinical implication: More rigorous RCTs specifically reporting modality parameters are urgently needed for severe stroke patients.

🤖 TIER 1–2 — Robot-Assisted Upper Limb Therapy (Multiple Reviews)

6. Robot-Assisted Therapy for Upper Limb — Umbrella Review (Stroke Journal, 2025)

Park JM et al. Stroke. 2025. DOI: 10.1161/STROKEAHA.124.048183

• Robot-assisted therapy significantly improves Fugl-Meyer Assessment (FMA) scores vs. conventional therapy (SMD 0.29) and as an add-on (SMD 0.42).
• ⚠️ Critical caveat: Improvements do NOT meet minimum clinically important difference (MCID) thresholds (12.4 for subacute; 3.5 for chronic stroke).
• Benefits for muscle strength (SMD 0.46); no significant effect on spasticity.
• Consistent effects regardless of stroke stage, robot type, or training site.

7. Robot-Assisted Therapy — Umbrella Review (ICF Framework) (J Med Internet Res, 2026)

Liu S et al. J Med Internet Res. 2026. DOI: 10.2196/79363

• RAT superior for upper limb motor function but limited evidence for ADL improvement.
• Structured using the International Classification of Functioning (ICF) framework — most comprehensive synthesis to date.
• Subgroup analysis: treatment effects vary by stroke stage, impairment severity, and robot type.
• 17 of 21 included reviews rated high quality (AMSTAR 2).

Clinical implication: Robotic therapy improves motor function at the body-structure level but struggles to translate to real-world ADL gains — personalisation by patient stage and impairment level is key.

⚡ TIER 3 — High-Profile RCTs

8. tDCS + Constraint-Induced Movement Therapy — TRANSPORT2 Trial (Lancet Neurology, 2025)

Schlaug G et al. Lancet Neurol. 2025. DOI: 10.1016/S1474-4422(25)00044-4

• Adding transcranial direct current stimulation (tDCS) at 2 mA or 4 mA to modified CIMT did not further reduce motor impairment vs. sham + CIMT at 15 days post-intervention.
• FMA improvement: sham+mCIMT = 4.91 pts; 2mA = 3.87 pts; 4mA = 5.53 pts (p=0.39 — not significant).
• tDCS was safe, well-tolerated, and feasible.
• Higher doses (>4 mA) may warrant future trials.

Clinical implication: tDCS does not augment CIMT outcomes at current standard doses — negative but important finding that refines combined modality protocols.

9. Lower Extremity–CIMT for Gait in Chronic Stroke (RCT, 2024)

Menezes-Oliveira E et al. Brain Inj. 2024.

• LE-CIMT improved gait and balance function in chronic post-stroke patients.
• Expands CIMT application beyond the upper limb.

10. Trunk Exercises with Hip Strategy Training (Multicenter RCT, 2024)

Almasoudi AO et al. Physiother Res Int. 2024.

• Trunk exercises combined with hip strategy training significantly improved independence and balance in stroke patients.
• Supports core-focused physiotherapy in stroke rehab programs.

11. Virtually Delivered Upper Limb Activity Program (Multicenter RCT, 2025)

Simpson LA et al. Neurorehabil Neural Repair. 2025.

• A virtually-delivered (telerehabilitation) program significantly increased upper limb activity post-stroke.
• Supports the growing evidence base for remote/digital physiotherapy delivery.

12. Blood Flow Restriction Training vs. Traditional Resistance Exercise (RCT, 2024)

Ahmed I et al. Top Stroke Rehabil. 2024.

• Low-intensity resistance training with blood flow restriction (BFR) comparable to traditional resistance exercise for lower extremity strength and motor function in ischemic stroke survivors — potentially safer for deconditioned patients.

📋 Summary Table

Key Themes Emerging from 2024–2026 Evidence

1. No single "best" approach — therapy dose, intensity, and task-specificity matter more than the named method (Cochrane 2025).
2. Robotic/electromechanical therapy improves motor function but clinical significance remains uncertain — best used as an add-on to conventional PT.
3. Virtual reality has growing evidence across acute to chronic phases, particularly for balance and upper limb.
4. tDCS combined with CIMT does not add benefit at standard doses — the neuromodulation + PT combination needs rethinking.
5. Telerehabilitation is gaining RCT-level evidence as a valid delivery model.
6. Severe stroke remains an evidence gap — modality parameters are poorly reported in existing studies.