研究課題情報
課題管理番号
25wm0625410h0002
統合プロジェクト(三期)

シーズ開発・基礎研究プロジェクト

統合プロジェクト(二期)

疾患基礎研究プロジェクト<第二期>

事業名

脳神経科学統合プログラム

タグ(2025)

研究開発目的: 研究開発基盤の整備研究

タグ(2024)

研究開発目的: 研究開発基盤の整備研究

研究の性格: AI・ICT等研究及びDX基盤整備研究

研究モダリティ: システム・機器の開発

開発フェーズ: 基盤構築

ICD-10大分類: 神経系の疾患

対象疾患領域: 精神・神経疾患 ,  その他

特記的事項: DX<SaMD、AI、ICTを含む>

補足確認: 予後・生活の質<QOL>

代表研究機関
学校法人沖縄科学技術大学院大学学園
研究代表者

(2025) GUTIERREZCarlos・Enrique, 学校法人沖縄科学技術大学院大学学園, 神経計算ユニット Visiting Researcher

(2024) GUTIERREZCarlos・Enrique, 学校法人沖縄科学技術大学院大学学園, 神経計算ユニット Visiting Researcher

研究期間
2024年度 - 2026年度
課題への総配分額
26,000 千円
研究概要(2025)
This research aims to develop BrainScaler, an AI-powered framework for multiscale brain modeling. BrainScaler will be an AI-powered tool oriented to harness data from Brain/MINDS and research publications. This objective involves several targets. Firstly, manual extraction of parameters from scientific publications is time-consuming. BrainScaler seeks to streamline this process by introducing an AI powered research scientist “bot”. This bot, based on fined-tuned large language models (LLMs), will aid in extracting brain model specifications from relevant publications datasets through natural language interactions. Furthermore, BrainScaler will incorporate multiscale modeling workflow, to generate models across different levels of organization. It will implement functions for modeling population of neurons, such as mean-field and rate models. Additionally, BrainScaler will ensure interoperability with both Brain/MINDS Marmoset Connectivity Resource and the Virtual Brain.
研究概要(2024)
This research aims to develop BrainScaler, an AI-powered framework for multiscale brain modeling. BrainScaler will be an AI-powered tool oriented to harness data from Brain/MINDS and research publications. This objective involves several targets. Firstly, manual extraction of parameters from scientific publications is time-consuming. BrainScaler seeks to streamline this process by introducing an AI powered research scientist “bot”. This bot, based on fined-tuned large language models (LLMs), will aid in extracting brain model specifications from relevant publications datasets through natural language interactions. Furthermore, BrainScaler will incorporate multiscale modeling workflow, to generate models across different levels of organization. It will implement functions for modeling population of neurons, such as mean-field and rate models. Additionally, BrainScaler will ensure interoperability with both Brain/MINDS Marmoset Connectivity Resource and the Virtual Brain.
研究成果情報
2024