Academic research: what exercises are effective for the shoulders (strength, mobility, common desk/posture issues)?

Generated by /academic-research on 2026-08-22. Synthesized across 3 rounds from 15 peer-reviewed papers (see Provenance). Treat as raw material — review before promoting into a project or thread. Context: vault/threads/shoulder-exercises Abstracts-only limitation (v1): claims below are synthesized from Consensus abstracts + metadata, not full papers. Consensus in this environment returned the top 3 hits per query (about 20 found per search; only those 3 abstracts were retrieved). Priors skipped (Paul not present). Not medical advice. This synthesis does not diagnose injury, does not treat rotator-cuff tears, and does not prescribe rehab. Clinical decisions stay open.

Summary

Retrieved abstracts support three kinds of work, not one best routine. Progressive resistance training improves strength and hypertrophy in healthy adults, with heavier loads (≥80% 1RM), 2–3 sets, and ≥2 sessions/week ranked for strength (Currier et al. 2026; Currier et al. 2023) — those overviews are whole-body, not a named shoulder-lift ranking. For the rotator-cuff / deltoid muscles specifically, one EMG study in healthy adults reports sidelying external rotation as the highest infraspinatus / teres minor activation among seven common ER drills, and prone horizontal abduction at 100° with full ER as the highest for supraspinatus and deltoid (Reinold et al. 2004). One preliminary isokinetic study reports eccentric ER training raised eccentric strength (~24%) and supraspinatus fascicle length (~16%) in physically active men (Vetter et al. 2023).

On mobility, sleeper stretching and posterior glenohumeral mobilization each increased internal-rotation and horizontal-adduction PROM in a single session (Swanson et al. 2024); cross-body stretching improved IR and horizontal adduction, more so when combined with dorsal-glide mobilization (Kang et al. 2019); a 7-week high-volume pectoralis-major stretch increased shoulder-extension ROM (~6%) and isometric torque at long muscle length (~11%) in healthy adults (Reiner et al. 2023).

On desk / posture, a 2024 systematic review and meta-analysis reports that therapeutic exercises (strength, stretching, shoulder-based, and “comprehensive” programs) reduced forward-head, rounded-shoulder, and thoracic-kyphosis angles in people labeled with upper-crossed syndrome (Sepehri et al. 2024). Office-worker RCTs report gains from scapular-stabilization plus thoracic-extension work (Kang et al. 2021), from online-supervised or workplace corrective exercise (Yaghoubitajani et al. 2022), and from adding shoulder-stabilization work to office exercises (Depreli et al. 2024). Combining pectoralis-minor stretching with lower-trapezius / scapular posterior-tilt work outperformed stretch-or-tilt alone on some posture and muscle-activation measures (Lee et al. 2015; Hasan et al. 2023).

No retrieved abstract names a single “best” gym list (overhead press, lateral raise, face pull, etc.). Thoracic mobility exercise tied to shoulder ROM appears in one tennis multimodal program (Le Solliec et al. 2023); adding a single-session thoracic manipulation before sleeper stretch or mobilization reduced IR gains (Swanson et al. 2024). Those are not the same intervention.

Findings

Strength: progressive loading works; named shoulder lifts are thin in these abstracts

An ACSM-commissioned overview of 137 systematic reviews (>30,000 participants) reports that resistance training versus no exercise improved muscle strength, size, power, endurance, and several function outcomes in healthy adults. Few prescription variables changed primary adaptations. Voluntary strength was enhanced by heavier loads (≥80% 1RM), a complete range of motion, 2–3 sets, placing the lift at the start of the session, and ≥2 sessions/week; hypertrophy by higher weekly volume (≥10 sets/week) and eccentric overload (Currier et al. 2026). A 2023 Bayesian network meta-analysis of the same first author (178 strength trials, n=5,097; 119 hypertrophy trials, n=3,364) ranks all resistance-training prescriptions above no-exercise control; the highest-ranked strength prescription was higher-load, multiset, thrice-weekly training (SMD 1.60 vs control), and the highest-ranked hypertrophy prescription was higher-load, multiset, twice-weekly (SMD 0.66 vs control) (Currier et al. 2023).

What this does not license. Those reviews synthesize whole-body resistance training. The retrieved abstracts do not name overhead press, lateral raise, seated row, or face pull as shoulder-specific winners. A query aimed at those lifts returned the same ACSM overview, not a deltoid-exercise bake-off.

Shoulder-specific strength evidence in this corpus is narrower. A preliminary six-week twice-weekly eccentric isokinetic external-rotation program in 16 physically active men reported +24% eccentric strength (p=.008), +16% supraspinatus fascicle length (p=.003), and +19% fascicle volume (p=.002), with a 4% decrease in passive ROM on a stretch test (Vetter et al. 2023). That is one small, equipment-specific study, not a gym-exercise ranking.

An intramuscular-EMG study in 10 healthy adults compared seven common external-rotation exercises. Sidelying ER at 0° abduction produced the greatest infraspinatus (62% MVIC) and teres minor (67% MVIC) activity. Prone horizontal abduction at 100° with full ER produced the greatest supraspinatus (82% MVIC), middle deltoid (87% MVIC), and posterior deltoid (88% MVIC) activity (Reinold et al. 2004). EMG is activation, not a measured strength gain. The abstract frames the work as information “to develop rehabilitation programs” and for future studies — not a treatment prescription.

Mobility: sleeper stretch, cross-body stretch, and pec stretching move ROM in the abstracts

In 40 people with clinically meaningful internal-rotation loss, a single session of either grade-III posterior glenohumeral mobilization or sleeper stretching improved IR and horizontal-adduction PROM within group; the two local interventions did not differ from each other. Adding a supine HVLA thoracic manipulation at T3–4 before those local drills produced smaller IR gains (mean difference 4.4° after mobilization, 6.4° after sleeper) (Swanson et al. 2024).

In 40 people with glenohumeral IR deficits, cross-body stretching alone increased IR, horizontal adduction, and a shoulder-mobility measure; combining it with simultaneous dorsal-glide mobilization produced larger changes (IR +6°, HA +10°, mobility −2 cm vs stretch alone) (Kang et al. 2019).

A 7-week, 3×/week, 15-minute pectoralis-major static-stretch program in 38 healthy, physically active adults increased shoulder-extension ROM (~6%, d=0.92) and isometric peak torque at a long muscle length (~11%, d=0.76), with no change in pec stiffness or torque at a short length, and no change in controls (Reiner et al. 2023). That is a healthy-adult mobility/strength finding, not a desk-posture trial.

Desk / posture: stretch plus scapular / thoracic work, not a single “office stretch”

A systematic review and meta-analysis of 22 studies on therapeutic exercise for upper-crossed syndrome reports significant reductions in forward-head, rounded-shoulder, and thoracic-kyphosis angles. The authors conclude that strength, stretching, shoulder-based, and especially comprehensive programs that target all involved muscles “may be effective” for those angles and for UCS overall (Sepehri et al. 2024). UCS is a label used in those papers, not a diagnosis issued here.

Office-worker trials in this corpus:

• An RCT in 32 office workers with forward-head posture compared scapular-stabilization plus thoracic-extension exercise (not applied to the cervical spine) with cervical stabilization plus stretching. Both groups improved craniovertebral angle, pain, and neck-disability scores; the scapular/thoracic group also improved several respiratory measures, and between-group differences favored that group on FEV1 and pain (Kang et al. 2021). Thoracic extension exercise is in scope because this abstract ties it to a shoulder/posture office-worker question (FHP plus scapular work).
• An RCT in 36 office workers meeting alignment and pain thresholds labeled as UCS compared 8-week online-supervised corrective exercise, workplace exercise, and usual-activity control. Both exercise groups improved neck-shoulder pain and photogrammetry angles (forward head, rounded shoulder, rounded back) from baseline; the online-supervised arm differed from control on pain, postural angles, workability, and upper-trapezius activation (Yaghoubitajani et al. 2022).
• In 60 office workers with shoulder protraction, adding 8 weeks of shoulder-stabilization exercise to office exercise improved protraction, more muscle-strength sites, proprioception, and closed-kinetic-chain performance than office exercise alone; some strength and performance gains also appeared in the office-exercise-only arm (Depreli et al. 2024).
• In 38 office workers, 6 weeks of scapular-stability exercise (40 min, 2×/week) improved forward-head distance, rounded-shoulder posture, upper-limb stability, and lower-trapezius thickness on ultrasound more than a manual-therapy comparator (Go et al. 2016).

On the stretch-plus-strengthen pairing for rounded-shoulder posture: in 15 young men with RSP, scapular posterior-tilt exercise after pectoralis-minor stretching, and the same tilt exercise with a shoulder brace, reduced RSP and increased a pec-minor length index more than tilt exercise alone; lower-trapezius EMG was highest after stretch-then-tilt (Lee et al. 2015). In 60 young Saudi women with rounded shoulders, combining lower-trapezius strengthening with pec-minor stretching increased pec-minor resting length more than stretching alone; shoulder-flexion ROM improved in both arms without a between-group difference (Hasan et al. 2023).

What this does not license. These trials measure posture angles, some pain/workability scores, and EMG — they are not a diagnosis of injury and not a rehab protocol for a tear. Several samples were selected on UCS / FHP / RSP thresholds. Abstracts do not prescribe a consumer workout app or a gear purchase (Lee’s brace arm is a lab comparison, not a product review).

Thoracic mobility is tied to the shoulder in two different ways — and they disagree

An 8-week multimodal program (stretching, strengthening, myofascial release, 4×/week) in 22 competitive young tennis players increased thoracic mobility, decreased interscapular distance, and increased glenohumeral internal and external rotation ROM, with moderate-to-strong increases in serve accuracy and velocity; the preceding 8 weeks of regular tennis training alone did not move those measures (Le Solliec et al. 2023). That abstract does tie thoracic work to a shoulder ROM outcome — so it stays in scope. It is one tennis-player study, not office workers.

Swanson et al. (already cited) found that adding a single-session thoracic manipulation reduced the IR-PROM gain from sleeper stretch or glenohumeral mobilization (Swanson et al. 2024). Thoracic extension exercise over weeks (Kang 2021; Le Solliec 2023) is not the same intervention as a one-session HVLA manipulation. Do not collapse them.

Contradictions and open questions

• No named “best” strength lift. Consensus queries for overhead press / lateral raise / seated row returned whole-body RT overviews, not a shoulder-exercise ranking. EMG (Reinold et al. 2004) names drills; it does not prove those drills produce the largest 1RM or hypertrophy change.
• Activation ≠ adaptation. Reinold reports %MVIC; Vetter reports isokinetic eccentric training on a dynamometer. Those are different outcomes and populations.
• Thoracic work is split. Multi-week thoracic-extension / multimodal programs improved FHP or GH ROM in the retrieved abstracts (Kang et al. 2021; Le Solliec et al. 2023). One-session thoracic manipulation reduced IR gains from local shoulder mobility work (Swanson et al. 2024).
• UCS / FHP / RSP are paper labels. Sepehri, Yaghoubitajani, Kang 2021, Lee, Hasan, Depreli, and Go enroll on those posture clusters. This file does not diagnose them.
• Pain / RCRSP literature was retrieved and set aside. Round-3 searches for shoulder-strengthening systematic reviews returned rotator-cuff-related pain meta-analyses (exercise reduced pain and improved function in middle-aged/older adults with RCRSP). Those abstracts are treatment-of-pain papers. They are not used as answers here, because this thread is not a diagnosis or rehab protocol.
• Abstracts omit protocols. Sepehri’s 22-study meta-analysis does not name the exercise lists in the abstract. Office-worker RCTs name classes of work (scapular stabilization, thoracic extension, pec-minor stretch, lower-trap strengthening) more often than set/rep schemes.
• Full text would be needed for effect-size tables, exact exercise descriptions, and whether “office exercises” overlap with the stabilization programs.

Provenance

Rounds run: 3 (full)

Sub-questions by round:

Round 1 (broad survey):

1. Which resistance-training exercises produce the greatest deltoid and rotator-cuff muscle activation or strength gains in healthy adults?
2. Which stretching or mobility interventions improve glenohumeral range of motion?
3. What exercise interventions address desk-work / sitting-related rounded-shoulder or forward-head posture at the shoulder?
4. How do scapular-stabilization exercises affect shoulder strength and posture in non-clinical office or general adult samples?

Round 2 (drill-down):

1. EMG comparison of named external-rotation and prone Y/T/W-style drills — targeting the missing named-exercise list from round 1
2. Pectoralis stretch plus scapular retraction / posterior tilt for rounded-shoulder posture — targeting the stretch-vs-strengthen pairing
3. (Round-1 Q4 search landed in the same batch.) Scapular stabilization in office workers with protraction — targeting desk-specific strength/posture

Round 3 (resolve remaining uncertainty):

1. Named compound lifts (overhead press, lateral raise, row) for shoulder strength/hypertrophy — targeting the still-missing gym-lift ranking
2. Systematic reviews of shoulder strengthening in healthy adults — targeting whether a shoulder-specific evidence synthesis exists without becoming a pain-rehab review
3. Thoracic mobility exercise effects on glenohumeral ROM — targeting SCOPE’s “thoracic only if papers tie it to the shoulder”

Papers reviewed (15 total; R1: 6, R2: 5, R3: 4):

Round 1:

• American College of Sports Medicine Position Stand. Resistance Training Prescription for Muscle Function, Hypertrophy, and Physical Performance in Healthy Adults: An Overview of Reviews — Currier et al. (2026). Medicine and Science in Sports and Exercise — SJR not returned, 33 citations, overview of reviews. Whole-body RT prescription; not a shoulder-lift ranking.
• Effect of isokinetic eccentric training on the human shoulder strength, flexibility, and muscle architecture in physically active men: A preliminary study — Vetter et al. (2023). PLOS ONE — SJR not returned, 13 citations, preliminary pre-post. Eccentric isokinetic ER; fascicle-length change.
• Effects of a High-Volume 7-Week Pectoralis Muscle Stretching Training on Muscle Function and Muscle Stiffness — Reiner et al. (2023). Sports Medicine - Open — SJR not returned, 18 citations, RCT. Pec-major stretch → extension ROM + long-length MVIC.
• The effect of various therapeutic exercises on forward head posture, rounded shoulder, and hyperkyphosis among people with upper crossed syndrome: a systematic review and meta-analysis — Sepehri et al. (2024). BMC Musculoskeletal Disorders — SJR not returned, 54 citations, systematic review and meta-analysis. Exercise reduced FHP / RSP / kyphosis angles.
• Effects of a combination of scapular stabilization and thoracic extension exercises for office workers with forward head posture on the craniovertebral angle, respiration, pain, and disability: A randomized-controlled trial — Kang et al. (2021). Turkish Journal of Physical Medicine and Rehabilitation — SJR not returned, 43 citations, RCT. Scapular + thoracic extension in office FHP.
• Corrective exercises administered online vs at the workplace for pain and function in the office workers with upper crossed syndrome: randomized controlled trial — Yaghoubitajani et al. (2022). International Archives of Occupational and Environmental Health — SJR not returned, 37 citations, RCT. Online-supervised vs workplace corrective exercise.

Round 2:

• Electromyographic analysis of the rotator cuff and deltoid musculature during common shoulder external rotation exercises. — Reinold et al. (2004). The Journal of orthopaedic and sports physical therapy — SJR not returned, 360 citations, repeated-measures EMG. Sidelying ER vs prone horizontal abduction ranking.
• The effect of scapular posterior tilt exercise, pectoralis minor stretching, and shoulder brace on scapular alignment and muscles activity in subjects with round-shoulder posture. — Lee et al. (2015). Journal of electromyography and kinesiology — SJR not returned, 75 citations, comparative intervention. Stretch-then-tilt beat tilt alone for RSP and LT EMG.
• The Combined Effect of the Trapezius Muscle Strengthening and Pectoralis Minor Muscle Stretching on Correcting the Rounded Shoulder Posture and Shoulder Flexion Range of Motion among Young Saudi Females: A Randomized Comparative Study — Hasan et al. (2023). Healthcare — SJR not returned, 17 citations, RCT. LT strengthen + PMi stretch vs stretch alone.
• The effects of shoulder stabilization exercises on muscle strength, proprioceptive sensory ability and performance in office workers with shoulder protraction — Depreli et al. (2024). International Journal of Occupational Safety and Ergonomics — SJR not returned, 3 citations, RCT. Stabilization added to office exercise.
• Effects of scapular stability exercise on shoulder stability and rehabilitative ultrasound images in office workers — Go et al. (2016). Journal of Physical Therapy Science — SJR not returned, 14 citations, RCT. Scapular-stability vs manual therapy.

Round 3:

• Resistance training prescription for muscle strength and hypertrophy in healthy adults: a systematic review and Bayesian network meta-analysis — Currier et al. (2023). British Journal of Sports Medicine — SJR not returned, 151 citations, systematic review and Bayesian NMA. Highest-ranked load/set/frequency combinations; still not shoulder-lift-specific.
• Combined Effects of Glenohumeral Mobilization, Stretching, and Thoracic Manipulation on Shoulder Internal Rotation Range of Motion — Swanson et al. (2024). International Journal of Sports Physical Therapy — SJR not returned, 1 citation, randomized sequential lab study. Sleeper vs PGM; thoracic manipulation reduced IR gains.
• Effects of self-stretching with mobilization on shoulder range of motion in individuals with glenohumeral internal rotation deficits: a randomized controlled trial. — Kang et al. (2019). Journal of shoulder and elbow surgery — SJR not returned, 8 citations, RCT. Cross-body stretch ± dorsal glide.
• Effects of an 8-week multimodal program on thoracic posture, glenohumeral range of motion and serve performance in competitive young tennis players — Le Solliec et al. (2023). Frontiers in Sports and Active Living — SJR not returned, 9 citations, pre-post vs regular-training baseline. Thoracic/shoulder multimodal program tied to GH ROM.

Papers seen and not retained (scope or quality):

• Barros et al. 2023 (PLOS ONE) — RCT protocol for isometric vs isotonic exercise in rotator-cuff tendinopathy; no results, and the population is a clinical diagnosis.
• Vico-Rodríguez et al. 2026, Zhang et al. 2025, Wu et al. 2025 — systematic reviews of exercise for rotator-cuff-related shoulder pain. Treatment-of-pain; not used as answers.
• Hwang et al. 2021 — PNF for office workers labeled with scapular dyskinesis (more clinical/rehab than the desk-exercise question).
• Fani et al. 2020 — scapular mobilization vs pec-minor stretch in RSP; overlapping with Lee/Hasan and over the paper cap.
• Mcleod et al. 2023 — umbrella review of RT variables; redundant with the two Currier papers.
• Antari et al. 2021 — frozen-shoulder mobilization; diagnosis/treatment, out of scope.

Tools used: mcp__consensus__search (Consensus — covers Semantic Scholar, PubMed, Scopus, ArXiv). Filters applied: none. Generated: 2026-08-22 13:30 UTC