Generate a Prompt for Interpreting Brain–Behavior Associations

Source: R/subtyper.R

Constructs a standardized prompt for guiding large language models in interpreting statistical associations between imaging-derived phenotypes (IDPs) and human performance measures.

generate_idp_interpretation_prompt(
  response_length = c("short", "medium", "long"),
  tone = c("critical", "skeptical", "neutral"),
  n_examples = 3
)

Arguments

response_length

Character. Desired length of justification. One of: "short" (2–3 sentences), "medium" (3–6), "long" (5–8).

tone

Character. Interpretive stance. One of: "neutral", "skeptical", "critical".

n_examples

Integer. Number of exemplar cases to include (0–5).

Value

A character string containing the full prompt.

Details

The model is required to return a strict JSON object with exactly three keys:

  • "consistency" : low / medium / high

  • "justification" : academic explanation

  • "plausibility" : continuous numeric score between zero and one 0.0–1.0