> For the complete documentation index, see [llms.txt](https://docs.neurolabai.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.neurolabai.io/examples-and-use-cases/ketamine-multisystem-investigation.md). # Explore a Multi-System Report Ketamine is the advanced breadth example in this guide. Its frozen response is rich enough to inspect mechanism, cognition, usage, safety, evidence provenance, and brain localization in one session. The lesson is not that a larger report is a better result. It is how to keep different claim types separate when the application returns a great deal of material. ## Start with the report summary The frozen beta example returned `NMDA receptor antagonist` first in the MoA head (score 1.017), with 3-of-3 seed agreement. The cognition head returned `depression` first (0.715), also with 3-of-3 agreement. ![Ketamine Chat Report showing the mechanism summary in the report interpreter](https://3908477117-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fgze1K0yofnhFNmM9nEyt%2Fuploads%2Fgit-blob-94da943d1fe63ff8f45e9c392e8d6b3af607b714%2F04-chat-report.jpg?alt=media) Read those as two ranked outputs from different heads. The cognition label must not be rewritten as a treatment recommendation or an efficacy claim. Scores across heads are not comparable. ## Separate agreement from coverage The MoA seeds agree, but the report also says: * 11 binding/MoA-relevant edges are available against a threshold of 21; * MoA coverage is `low_kg_coverage`; * the receptor-to-brain decoder has 10 informative receptor hits and is not decision-grade; and * nucleus coverage is `bridge_only`, with no curated Compound-to-Nucleus edge. Agreement asks whether the ensemble members were consistent. Coverage asks whether the KG supplied the kind and amount of context expected by the release criteria. Both belong in the interpretation. ## Inspect the evidence composition The response's high-level evidence summary contains 1,184 rows, but 1,146 are classified as uncited. The report-local KG bundle is deliberately smaller: 98 nodes, 109 edges, 109 paths, and 14 sources. ![Ketamine report-local evidence graph centered on the compound and its returned entities](https://3908477117-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fgze1K0yofnhFNmM9nEyt%2Fuploads%2Fgit-blob-7f7439286584fe7b75991e9cae6894a4ff4dfd14%2F07-knowledge-graph-detail.jpg?alt=media) This is why counts should never stand alone. Ask: 1. Is the record observed, inferred, derived, or unknown? 2. Is its source directly surfaceable, redacted, training-only, or unavailable? 3. Does a support link point to this exact field? 4. Is the graph showing query-specific support or optional reference context? An evidence artifact can be complete relative to its declared report scope while the scientific literature and provenance coverage remain incomplete. ## Treat label-derived phrases as extracted records The frozen `usage.fda_labeled` array contains label-linked phrases, including `anesthetic agent`, alongside broad fragments. A link to an FDA/DailyMed record identifies provenance. It does not certify that every extracted phrase is a clean approved-indication label. Before reusing a phrase, open the source and verify its section and meaning. Keep the source-derived array separate from `reported_usage_top_k`, which the release itself marks `weak_non_cns_only` and not authoritative for CNS compounds. ## Keep the localization boundary visible The captured Brain Explorer defaults to **Legacy specific map (diagnostic)**. Its strongest displayed parcels and network summaries are relative values from this diagnostic channel. ![Ketamine Brain Explorer focused on the interactive brain and tract rendering](https://3908477117-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fgze1K0yofnhFNmM9nEyt%2Fuploads%2Fgit-blob-fe73590e3122436c84efcb698b7a3fb3cdb4ceba%2F08-brain-localization-detail.jpg?alt=media) Do not describe the display as measured activation, a patient scan, a validated compound-specific BOLD map, or evidence of causal propagation. In this report, the absence of curated nucleus edges is an additional reason to avoid turning a bridge-derived shortlist into a top-1 biological claim. ## A defensible summary > In the pinned v43.262 release, Neurolab ranked NMDA receptor antagonist first for Ketamine with unanimous seed agreement. The report supplies a broad report-local evidence graph, but MoA coverage remains below threshold, most evidence-summary rows are uncited, the CNS reported-usage head is non-authoritative, and brain/nucleus outputs remain exploratory. *** **Fixture:** sanitized SHA-256 `ab6ef26f478cd796a5bc0393f4453b72480b7f3f835b10c28b3cb722b2154a04`. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.neurolabai.io/examples-and-use-cases/ketamine-multisystem-investigation.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.