# Study 000D Manuscript ## Title Selective Expression Versus Normalization In A Six-Year Altered-Mechanics Adaptation Program ## Core Question Did the six-year system adapt through broad normalization, or through selective expression? ## Short Answer The corrected full program supports a `selective adaptation model` in which successful running expression became narrower, more stabilized, more turnover-dependent, and only partially less costly, rather than broadly normalizing across contexts. ## Program Role This package does not replace `Study 000A`, `Study 000B`, `Microstudy A`, `Microstudy B`, or `Study 000C`. Its role is narrower and more direct: - treat the existing study program as the evidence base - define two competing adaptation signatures - test which signature the corrected full dataset supports more strongly The result is not just a comparison exercise. It is a program-level thesis test. ## Competing Signatures ### Normalization signature If adaptation were trending toward broad normalization, the later system would be expected to show: - broader successful running expression across environments - less dependence on stabilized context - speed gain distributed more evenly across multiple mechanical degrees of freedom - better preservation of stride expression under higher-demand probes - disappearance of unexplained session-level burden ### Selective-expression signature If adaptation were selective rather than broadly normalizing, the later system would be expected to show: - increasing concentration of successful running expression in selected conditions - heavier dependence on stabilized context - speed gain assembled mainly through the most accessible lever - preserved turnover with more constrained stride expression under higher-demand probes - improved efficiency without full disappearance of burden ## Methods This study used only packaged tables from the already completed study program: - `Study 000A` - `Study 000B` - `Microstudy B` - `Study 000C` - the yearly ecology source table bundled in `Study 000A` The test was structured around six evidence families: 1. ecological breadth 2. mechanical gain pathway 3. higher-demand probe behavior 4. constraint persistence 5. internal cost 6. robustness after the full-data correction No opaque composite score was used. Each evidence family was evaluated directly against the two competing signatures. ## Results All six evidence families aligned more strongly with `selective expression`. ### 1. Ecological breadth favored selective expression The later system did not broaden outdoor running expression. Instead, successful running became more concentrated in stabilized context. The clearest late contrast was: - `2025` running share of structured hours: `21.00%` - `2026` running share of structured hours: `84.18%` - `2025` treadmill share of running miles: `92.78%` - `2026` treadmill share of running miles: `98.74%` - `2025` outdoor share of running miles: `7.22%` - `2026` outdoor share of running miles: `1.26%` This does not read like broadening environmental tolerance. It reads like narrowed successful expression. ### 2. Mechanical gain pathway favored selective expression Later speed gain was not distributed evenly. It was assembled mainly through turnover. In the first-versus-last-30 QC-pass run comparison: - speed gain: `36.11%` - cadence change: `21.15%` - stride-length change: `12.24%` - cadence share of cadence-stride speed gain: `62.42%` That pattern supports a system using the more accessible lever, not one remodeling uniformly. ### 3. Higher-demand probe behavior favored selective expression The later outdoor probe set did not preserve stride expression near stabilized-context expectations. Instead, turnover held up better than stride. In the later specialized outdoor subset: - cadence residual: `+11.15%` - stride residual: `-10.58%` - cadence above expected: `5/5` - stride below expected: `5/5` - vertical ratio above expected: `5/5` That is one of the strongest pieces of evidence in the whole program. ### 4. Constraint persistence favored selective expression Constraint-like signals did not dissolve as the program matured. Two facts mattered most: - `vertical_ratio_pct` was the lowest-variability mechanics candidate in `Study 000A` - in `Microstudy B`, vertical ratio stayed above expected in all `8/8` outdoor QC-pass runs and `5/5` later specialized outdoor runs The program therefore supports persistence of an anchored signal across context changes, not disappearance of constraint. ### 5. Internal cost favored selective expression Efficiency improved, but unexplained session-level burden did not disappear. In `Study 000B`: - speed-per-HR rose from `0.01559` to `0.01768` - HR residual shifted from `-1.59691 bpm` to `3.09257 bpm` In `Microstudy B`: - later outdoor mean HR residual: `8.67%` So the later system became more effective, but not simply cheaper in all contexts. ### 6. Full-data correction favored selective expression `Study 000C` corrected a real ecological issue: - the early flagship window included `7` GPS-bearing hybrid cardio sessions hidden under Garmin `indoor_cardio` But the correction did not overturn the main model. It retained: - phase-structured adaptation - turnover-led gain - conserved mechanics - mixed internal cost That is exactly what a robust selective-expression model should survive. ## Primary Conclusion The strongest supported reading is: `this altered-mechanics system adapted through selective expression rather than through broad normalization` More specifically: - ecological expression narrowed rather than broadened - stabilized running context became more central rather than less - cadence remained the more accessible adaptation lever - stride expression remained more constrained under higher-demand probes - burden improved only partially rather than disappearing Put more directly: `the program supports a selective adaptation model in which successful running expression became narrower, more stabilized, more turnover-dependent, and only partially less costly` ## What This Study Means The contribution here is not a universal rule. The contribution is a stronger within-dataset conclusion: `successful adaptation in this system appears to have come from becoming more selective about where and how running could be expressed` That is a materially stronger answer than: - things changed - one metric stayed stable - treadmill was involved It reframes the program around a broader adaptation principle. ## What This Study Cannot Prove This study does not prove: - how a normative comparison system would adapt under the same conditions - the exact tissue-level mechanism behind selective expression - that the pattern automatically generalizes to all constrained movers - that treadmill versus outdoor fully maps the entire stabilization-demand ladder ## Bottom Line Within this six-year corrected program, the data supports a `selective adaptation model` more strongly than a `broad normalization model`.