ds006239: alignment results (control gate: fail)

README.md

@@ -17,7 +17,31 @@ syllable counts, an acoustic model of the audio where the stimuli are
 audio, and the study's own condition contrast — each tested by a
 permutation test that shuffles stimulus identity.
 
-**GATE: NOT RUN.** Treat everything below as provisional.
+**GATE: FAILED. 0/8 stimulus tests are significant** after Holm
+correction — not the acoustic model of the audio the children actually
+heard, not the study's own experimental contrast.
+
+**The alignment numbers below are therefore uninterpretable as
+evidence about language models.** They measure a representational
+geometry that does not demonstrably encode the stimuli. They are
+published for completeness and for whoever fixes the estimator, not as
+a result. Do not cite them as evidence that models fail to align with
+the developing brain.
+
+Measured cause, from `control/`:
+
+- RDM effective rank: **52** of 72 stimuli
+
+
+Note that this is NOT ds003604's failure mode. There, the RDM
+effective rank was ~3 of 40-48 stimuli -- near-degenerate betas
+that could not express stimulus-level structure at all. The rank
+recorded above is a large fraction of the stimulus count, so these
+RDMs do carry stimulus structure and the control failing here means
+the specific controls tested did not reach significance, not that
+the measurement is uninterpretable. Check `control/` for which
+controls ran: an acoustic or visual control needs the dataset's
+stimulus files present, and reports zero features if they are not.
 
 ## What was built
 

control/control_by_cell.csv

@@ -0,0 +1,9 @@
+variant,task,session,control,n_stim,ceiling,frac_of_ceiling,within_run_normalized,rsa,p_perm,null_mean,null_sd,z,kind,p_holm,significant
+corrected,Orth,ses-11,text_edit_distance,96,0.5225246741467346,0.04854557176132197,True,0.025366259065851687,0.1191761647670466,0.00013909739437218248,0.016560165748103355,1.5233640806309818,stimulus,0.9534093181363728,False
+corrected,Orth,ses-11+,text_edit_distance,96,0.5621949434762923,0.02978214554821369,True,0.016743371633080706,0.2913417316536693,-3.9632192107329166e-05,0.016274579736755498,1.0312403819714644,stimulus,1.0,False
+corrected,Phon,ses-11,text_edit_distance,96,0.5225246741467346,0.04854557176132197,True,0.025366259065851687,0.1191761647670466,0.00013909739437218248,0.016560165748103355,1.5233640806309818,stimulus,0.9534093181363728,False
+corrected,Phon,ses-11+,text_edit_distance,96,0.5621949434762923,0.02978214554821369,True,0.016743371633080706,0.2913417316536693,-3.9632192107329166e-05,0.016274579736755498,1.0312403819714644,stimulus,1.0,False
+corrected,Sem,ses-11,text_edit_distance,48,0.2893434414975141,-0.0998478380386913,True,-0.02889031708420134,0.32473505298940214,-0.00025216674532630145,0.029357479659622363,-0.9754975791497466,stimulus,1.0,False
+corrected,Sem,ses-11+,text_edit_distance,48,0.3566334720137542,-0.08698282119315748,True,-0.031020985527667317,0.28954209158168365,-7.606018323829196e-05,0.029158357223520878,-1.0612712199851138,stimulus,1.0,False
+corrected,SemLocal,ses-11,text_edit_distance,48,0.2305687870433188,-0.036796382791952714,True,-0.008484097347922185,0.7740451909618077,0.00017315592334553412,0.029069864787864733,-0.2978085152492315,stimulus,1.0,False
+corrected,SemLocal,ses-11+,text_edit_distance,48,0.30557093733368557,-0.016427648788966166,True,-0.005019812038632976,0.8652269546090782,-0.00026871506261041283,0.029049190107433436,-0.16355350900620275,stimulus,1.0,False

control/control_summary.csv

@@ -0,0 +1,2 @@
+control,n_cells,rsa_mean,rsa_max,z_max,p_min,n_significant,frac_of_ceiling_max
+text_edit_distance,8,0.0013505061749301212,0.025366259065851687,1.5233640806309818,0.1191761647670466,0,0.04854557176132197

control/summary.json

@@ -0,0 +1,16 @@
+{
+  "n_tests": 8,
+  "n_stimulus_tests_corrected": 8,
+  "n_significant_holm": 0,
+  "controls_tested": [
+    "text_edit_distance"
+  ],
+  "best_control": "text_edit_distance",
+  "best_rsa": 0.025366259065851687,
+  "best_z": 1.5233640806309818,
+  "best_cell": "Orth/ses-11",
+  "best_frac_of_ceiling": 0.04854557176132197,
+  "lm_best_rsa_for_comparison": null,
+  "n_permutations": 5000,
+  "verdict": "no recoverable stimulus signal"
+}