A BIOPHYSICAL NETWORK MODEL REVEALS THE LINK BETWEEN DEFICIENT INHIBITORY COGNITIVE CONTROL AND MAJOR NEUROTRANSMITTER AND NEURAL CONNECTIVITY HYPOTHESES IN SCHIZOPHRENIA

The paper fully specifies the modeling framework (PFC LIF network with small‑world topology and SC planned/reactive inputs), the fitted parameter vector r=(r1,r2,r3,p), experimental processing, and key fitted outcomes: for healthy controls the model achieves an error rate ≈34% with a Kolmogorov–Smirnov p-value 0.77, and for patients an error rate ≈44% with p-value 0.543; it also reports the calibrated parameters and CIs (e.g., HC: r1=2.135, r2=0.631, r3=1.33, p≈0.0101; SZ: r1=2.456, r2=0.885, r3=0.957, p≈0.02) and gives core network sizes and trial counts (N1=75 pyramidal, N2=25 interneurons; N3=31 SC neurons; N=100 trials) as well as the SC input equations and sloper distribution, which align closely with the candidate’s plan . However, the paper omits several numerical constants needed for exact replication (e.g., WS base degree k, many biophysical constants, and delays tdelayr/tdelayp), and the KS comparison is performed on group-averaged CDFs rather than raw samples, limiting strict inferential validity . The candidate solution outlines a correct, paper-consistent pipeline but declines to compute results, citing missing data/constants; this is partly justified (several key constants are indeed unspecified), yet some items they flagged as missing are present in the paper (e.g., network sizes, trial count, and sloper distribution). Hence, the paper is reproducibility-incomplete and the model submission is execution-incomplete.