@article{oai:ipsj.ixsq.nii.ac.jp:00210352,
 author = {藤原, 敏輝 and 鈴木, 聡 and Toshiki, Fujiwara and Satoshi, Suzuki},
 issue = {3},
 journal = {情報処理学会論文誌},
 month = {Mar},
 note = {本稿では非侵襲的脳機能計測である近赤外線分光法（NIRS）を用いた認知負荷の強度推定アルゴリズムを提案し，そのパラメータ調整法について述べる．本アルゴリズムでは，従来のNIRS認知負荷推定相当であるヘモグロビン（Hb）濃度変化要因に加え，その周波数要因の2つを用いて負荷強度を推定する．後者要因は，Hb濃度変化信号の統計的性質の周波数特性変化に着目したもので，その統計量の瞬時値を自己組織化粒子フィルタで推定する．そして同質的に難易度が異なる計算課題を課したときの前頭葉Hb濃度変化の計測データを用いて，提案する推定アルゴリズムが個人間差異が少なく有用となるよう調整パラメータの最良化を図った．その後，認知負荷強度の推定値と，設定課題難易度および難易量（個々人のタスク難易度を反映していると思われる回答時間）との無相関検定を行ったところ，実験有効協力者12人全員から有意に相関があることが確認でき（ピアソンの積率相関：N = 48，α = .05），提案アルゴリズム・調整法の有効性を示せた．, In this paper, we proposed an algorithm for estimating cognitive load intensity using near-infrared spectroscopy (NIRS) and its parameter adjustment method. This algorithm estimates the load intensity by combining two factors: the frequency characteristic factor, and the change in hemoglobin concentration (ΔHb) factor which is equivalent to the conventional NIRS cognitive load estimation. The former frequency factor focuses on the frequency characteristic change of the statistical property of ΔHb signal, and the instantaneous value of the statistic is estimated by a self-organizing particle filter. For reduction of interpersonal differences, we searched for the adequate parameters in the proposed algorithm by utilizing ΔHb data measured at the frontal lobe during calculation task with different difficulty levels. After that, a Pearson correlation test was performed between the estimated cognitive load intensities and the specified task difficulty levels, and between the estimates and the answer times that seemed to reflect the task difficulty of each individual, respectively. As a result, it was confirmed that there was a significant correlation (the Pearson product-moment correlation: N = 48, α = .05) from all 12 participants, demonstrating the effectiveness of the proposed algorithm.},
 pages = {860--866},
 title = {自己組織化粒子フィルタを用いたNIRS認知負荷強度推定の個人間差異低減パラメータ調整},
 volume = {62},
 year = {2021}
}