Music interventions in 132 healthy older adults enhance cerebellar grey matter and auditory working memory, despite general brain atrophy

The researchers compared two music-based interventions in healthy older adults:

• Piano practice group (experimental): Participants received weekly 60-minute group piano lessons and were asked to practice at home for at least 30 minutes per day, 5 days per week. They learned to play simple melodies, scales, and chord progressions on a keyboard.
• Musical culture group (active control): Participants attended weekly 60-minute group sessions focused on active music listening, music theory, and music history. They listened to excerpts, discussed musical structure, and completed listening exercises (e.g., identifying instruments, rhythms, or emotions in pieces). No instrument playing was involved.

Both interventions lasted 6 months. The primary outcome was grey matter volume change measured via structural MRI (voxel-based morphometry). Secondary outcomes included auditory working memory (measured with a tonal working memory task) and verbal working memory (measured with the Digit Span Backward test).

The study also tracked practice intensity (minutes per week of home practice or listening) and sleep quality (Pittsburgh Sleep Quality Index) as potential moderators.

• 132 healthy older adults (mean age = 68.4 years, range 60–79)
• Recruited from the community in Hannover, Germany
• All were right-handed, native German speakers
• Had no prior formal music training (less than 1 year of lessons in their lifetime)
• No neurological or psychiatric disorders, no hearing impairment, no MRI contraindications
• Normal cognitive function (Mini-Mental State Examination score > 26)
• 67 participants were randomly assigned to piano practice, 65 to musical culture
• 18 participants dropped out (9 per group), leaving 114 completers (58 piano, 56 musical culture)

• Grey matter volume: Structural T1-weighted MRI scans at baseline and after 6 months. Voxel-based morphometry (VBM) was used to compare grey matter density changes across the whole brain and in pre-specified regions of interest (hippocampus, primary auditory cortex, cerebellum).
• Auditory working memory: A tonal working memory task where participants heard sequences of tones and had to identify whether a probe tone was present in the sequence. Outcome was accuracy (percentage correct) and reaction time.
• Verbal working memory: Digit Span Backward test from the Wechsler Adult Intelligence Scale (WAIS-IV). Participants repeated sequences of digits in reverse order. Outcome was the longest span correctly recalled.
• Practice intensity: Self-reported daily logs of minutes spent practicing piano or listening to music at home.
• Sleep quality: Pittsburgh Sleep Quality Index (PSQI), a self-report questionnaire (0–21 scale, lower = better sleep).
• General cognitive status: Mini-Mental State Examination (MMSE) at screening to exclude dementia.

Design: This was a randomized controlled trial (RCT) with two parallel groups. Participants were randomly assigned to either piano practice or musical culture using a computer-generated randomisation list. The study was single-blind — the MRI analysts and cognitive testers were blinded to group assignment, but participants obviously knew which intervention they received (no sham control).

Duration: The intervention lasted 6 months, with MRI scans and cognitive tests at baseline and post-intervention. This is a relatively long duration for a brain plasticity study, which is a strength.

Statistical approach: The primary analysis used whole-brain voxel-based morphometry with a cluster-level family-wise error (FWE) correction at p < 0.05. They also performed region-of-interest (ROI) analyses in the hippocampus, primary auditory cortex, and cerebellum using small-volume correction. Cognitive outcomes were analysed with mixed ANOVAs (group × time). They also ran multiple regression models to test whether practice intensity and sleep quality predicted brain changes and cognitive improvements.

What this design can prove:

• Because it's an RCT with random assignment, it can establish causality — the music interventions caused the observed brain and cognitive changes (compared to what would have happened without any intervention).
• The active control group (musical culture) is a strong design feature because it controls for social engagement, group interaction, and general cognitive stimulation. Any differences between groups can be attributed specifically to playing an instrument versus listening to music.
• The 6-month duration allows detection of structural brain changes, which typically require weeks to months of consistent training.

What this design cannot prove:

• Without a no-intervention control group, the study cannot tell us whether any music intervention is better than doing nothing. Both groups showed improvements in some measures, but we don't know if these exceed natural fluctuations or placebo effects.
• The lack of blinding for participants means that expectations could influence effort, practice compliance, and self-report measures (like sleep quality and practice logs).
• The study cannot separate the effects of music per se from the effects of structured group activity — both groups received weekly social interaction and cognitive engagement.
• The sample is relatively homogeneous (German, healthy, well-educated, right-handed), limiting generalisability.

Major methodological weaknesses:

• No passive control group (e.g., waitlist or no-intervention)
• Self-reported practice logs (prone to overestimation and recall bias)
• High dropout rate (13.6%), though balanced between groups
• Only two time points (pre/post) — cannot track the trajectory of change

Primary outcome — Grey matter volume changes (whole-brain analysis):

• When both groups were combined, there was a significant increase in grey matter volume in three clusters:
  • Caudate nucleus (bilateral): cluster size = 1,248 voxels, peak t = 4.82, p(FWE-corrected) = 0.001
  • Rolandic operculum (right): cluster size = 1,088 voxels, peak t = 4.68, p(FWE-corrected) = 0.003
  • Inferior cerebellum (left lobule VIII): cluster size = 1,024 voxels, peak t = 4.55, p(FWE-corrected) = 0.005
• No significant group differences in whole-brain grey matter change — both interventions produced similar increases in these regions.

Region-of-interest analyses:

• Right primary auditory cortex: Significant group × time interaction (p = 0.04). Grey matter volume decreased in the musical culture group (mean change = −0.8%) but remained stable in the piano group (mean change = +0.1%). This suggests piano practice protected against atrophy in this region.
• Hippocampus: No significant changes in either group.
• Cerebellum (ROI): Grey matter increase was significantly associated with tonal working memory improvement (r = 0.28, p = 0.008) and with practice intensity (r = 0.22, p = 0.04).

Secondary outcomes — Cognitive performance:

• Tonal working memory: Both groups improved from pre to post (main effect of time: F(1,112) = 6.82, p = 0.01, η²p = 0.06). No group difference. The improvement was positively associated with cerebellar grey matter increase (β = 0.24, p = 0.02) and with sleep quality (better sleep → more improvement, β = 0.21, p = 0.04).
• Digit Span Backward (verbal working memory): Both groups improved (main effect of time: F(1,112) = 5.91, p = 0.02, η²p = 0.05). No group difference. Mean increase was approximately 0.3 digits (from ~6.1 to ~6.4).
• Practice intensity: Piano group practiced a mean of 187 minutes/week (SD = 98); musical culture group listened for a mean of 142 minutes/week (SD = 76). Higher practice intensity was associated with greater cerebellar grey matter increase (r = 0.22, p = 0.04) and greater tonal working memory improvement (r = 0.25, p = 0.02).

Whole-brain atrophy pattern:

• Despite the localised increases, the study also found a widespread pattern of grey matter decrease over 6 months consistent with normal aging. This included the frontal, temporal, and parietal cortices, and the thalamus. The average whole-brain grey matter loss was approximately 0.3–0.5% over 6 months, which matches the expected rate for this age group.

• Cerebellar grey matter increase: The clusters showing significant increases were modest in size (roughly 1,000 voxels, or about 1–2 cm³). This is a small but detectable change — comparable to the volume of a pea.
• Tonal working memory improvement: The effect size was small-to-medium (η²p = 0.06, meaning about 6% of the variance in improvement was explained by time). In practical terms, accuracy improved by about 3–5 percentage points (from ~78% to ~82% correct).
• Digit Span Backward improvement: An increase of 0.3 digits is a very small effect — roughly equivalent to being able to hold one extra digit in mind about 30% of the time.
• Practice intensity effect: For every additional 100 minutes of practice per week, cerebellar grey matter increased by about 0.1% (estimated from the regression slope). This is a dose-response relationship, but the effect is small.
• Atrophy rate: The whole-brain grey matter loss of 0.3–0.5% over 6 months translates to about 0.6–1.0% per year, which is the expected rate for healthy 68-year-olds. The music interventions did not stop this general atrophy — they only produced localised increases in specific regions.

Acknowledged by authors:

• No passive control group — cannot rule out that any structured activity would produce similar effects
• Self-reported practice logs may be inaccurate
• Single-blind design only (participants knew their group)
• Relatively short duration (6 months) for detecting brain changes
• Only two time points (cannot model trajectories)

Additional critical observations:

• The sample was highly selected: healthy, well-educated, no cognitive impairment, no hearing loss. Results may not generalise to less healthy or more diverse populations.
• The "musical culture" group was not a true control — it was an active intervention that also involved cognitive engagement and social interaction. The lack of group differences suggests that any structured music engagement (playing or listening) may be beneficial, but we cannot separate music from general cognitive stimulation.
• The primary outcome (whole-brain grey matter change) showed no group differences — the only group difference was in a secondary ROI analysis (right auditory cortex). This weakens the claim that piano practice is specifically superior.
• The association between sleep quality and cognitive improvement is correlational — better sleepers may have been more engaged or healthier overall.
• Practice intensity was self-selected (participants chose how much to practice), so the dose-response relationship could be confounded by motivation, baseline ability, or other factors.
• The study was funded by the Swiss National Science Foundation and the German Research Foundation — no obvious industry funding, but no mention of preregistration.

For someone running their own n=1 experiment:

What to test:

• Option A (piano practice): Learn to play simple melodies on a keyboard or digital piano. Aim for 30–60 minutes per day, 5–6 days per week. Focus on learning new pieces and scales (not just repeating familiar songs).
• Option B (active music listening): Listen to unfamiliar classical or instrumental music for 30–60 minutes per day. Actively engage by identifying instruments, following melodic lines, or analysing structure (not just background listening).
• Option C (combined): Alternate between playing and listening on different days.

Minimum meaningful duration:

• At least 3 months to see any detectable brain changes; 6 months is more reliable based on this study. Brain plasticity in older adults is slower than in younger people.
• Measure at baseline, 3 months, and 6 months to track trajectory.

What to measure:

• Primary metric: Working memory — use a free online digit span test (forward and backward) or a tonal memory test (e.g., from music aptitude tests). Test weekly at the same time of day.
• Secondary metric: Practice time — log minutes per day in a spreadsheet or app. Be honest — underreporting is common.
• Tertiary metric: Sleep quality — use the Pittsburgh Sleep Quality Index (free online) or a sleep tracker. Measure weekly.
• Optional: Subjective cognitive function — use a simple 1–10 scale for "how sharp did your memory feel today?"

Key confounds to control for:

• Social engagement: Both groups in the study involved weekly group sessions. If you do this alone, you're missing the social component. Consider joining a class or group, or at least playing/listening with a partner.
• Cognitive stimulation: Any new learning (language, chess, dancing) might produce similar effects. If you want to test music specifically, avoid starting other major cognitive hobbies during the experiment.
• Practice consistency: Sporadic practice (e.g., 4 hours on Saturday, nothing during the week) is less effective than daily 30-minute sessions. Aim for consistency over intensity.
• Age and baseline: If you're under 60, you may see faster changes. If you have prior music training, the effects may be smaller (ceiling effect).
• Hearing: Ensure you have normal hearing or use hearing aids. Hearing loss reduces the auditory stimulation needed for brain changes.
• Sleep: Poor sleep blunts brain plasticity. If your sleep is poor, fix that first before starting the experiment.

What a positive result would look like:

• Digit span backward: An increase of 1 or more digits (e.g., from 5 to 6) over 6 months. A 0.3-digit increase (as in the study) is barely noticeable — aim for a larger effect with consistent daily practice.
• Tonal memory: Improvement of 5–10 percentage points in accuracy on a tonal memory test (e.g., from 75% to 85% correct).
• Subjective: Feeling that you can hold more information in mind during conversations, or that you remember song melodies and lyrics more easily.
• Dose-response: The more you practice (especially >150 minutes/week), the larger the improvement. If you practice 30 minutes/day and see no change after 3 months, increase to 45–60 minutes/day.
• Caveat: You won't be able to measure grey matter changes at home. The cognitive tests are your proxy. If working memory improves, it's reasonable to infer that brain changes are occurring, based on this study's findings.