One bit per site gave sound match scores above chance in selected panels from six SEEG patients and 24 fMRI adults

We tested whether one bit per sampled site keeps within-group sound match accuracy in separate human neural studies. Public data supplied three SEEG high-gamma repeats per sound and three BOLD fMRI estimates per sound. The reader knew the broad sound group and made a two-choice template judgment. Calibration used other sounds. Each score averaged 32 fixed draws of 16 sites. Development used two patients and four MRI adults. A local code lock preceded tests in six other patients, 16 B2021 adults and eight NH2015 adults. Mean binary accuracies were 0.66983, 0.64829 and 0.63361. Their pointwise 95 percent person t intervals were [0.64743,0.69223], [0.63256,0.66402] and [0.61140,0.65582]. All 30 people were above chance. The fixed shared mean and majority rule passed after two-claim Holm correction, p=0.03125. The full-value reader did better in every person. The fixed 80 percent retention claim failed, p=1. Ratios of cohort mean advantages were about 76 to 77 percent. These are reader scores, not fractions of neural information. Eight-site and group-omission checks kept positive mean binary advantages. Some crossed sound resamples were undefined and remain so. An exact model with one scalar and mixed site loadings gave 0.77285 for both readers. Thus high match accuracy does not require a separate code for each sound or more than one latent signal dimension. Artificial same-data site selection gave mean binary accuracy 0.53659 with no stable sound signal. This is not a fitted estimate of source bias. The package gives code, source pins and terms, all outcomes, local plans, figures and repeat instructions. Original neural arrays stay external. Site selection, clinical sampling, child and adult cohorts, different source filters, known broad groups and fixed sounds limit scope. This is a new endpoint in reused studies, not a new acquisition. It does not establish a binary neural code, a common neural cause or causality.

human neuroscience · SEEG · fMRI · natural sounds · binary codes · quantization · repeat responses · sound matching