Authors: Sachio Otsuka, Megumi Nishiyama, Fumitaka Nakahara, & Jun
Kawaguchi(大塚幸生・西山めぐみ・中原史隆・川口 潤)
Title: Visual statistical learning based on the perceptual and semantic
information of objects
Journal(書誌情報): Journal of Experimental Psychology: Learning,
Memory, and Cognition, 39(1), 196-207, 2013.
doi: 10.1037/a0028645
論文URL: http://psycnet.apa.org/journals/xlm/39/1/196.pdf
Abstract: Five experiments examined what is learned based on the
perceptual and semantic information of objects in visual statistical
learning (VSL). In the familiarization phase, participants viewed a
sequence of line drawings and detected repetitions of various objects.
In a subsequent test phase, they watched 2 test sequences (statistically
related triplets vs. unrelated foils) and decided whether the first or
second sequence was more familiar based on the familiarization phase. In
Experiment 1A, the test sequences comprised line drawings; in Experiment
1B, they comprised word stimuli representing each line drawing. The
results showed that performance for statistically related triplets was
greater than chance. In Experiments 2 and 3 containing the forward ABC
and backward CBA triplets in the test, the results showed the importance
of temporal order, especially in line drawings. In Experiment 4, in
which the forward triplets were pitted against the backward triplets, we
showed that temporal order is still important for the expression of VSL
with word stimuli. Finally, in Experiment 5, we replicated the results
of Experiments 2 and 3 even with the images of visual objects. These
results suggest the parallel processes on the visual features and
semantic information of objects in VSL.
著者Contact先の email: otsuka.sachio.65w@st.kyoto-u.ac.jp
日本語によるコメント(オプション,200-300字で)
 本研究では,人がどのような情報に基づいて統計的規則性を抽出・学習してい
るのかを検討した。実験の結果,テスト時にオブジェクトが用いられた場合,学
習―テスト間で刺激列内のオブジェクトの呈示順序が同じ条件のみで学習効果が
見られ,順序が異なる条件では学習効果は認められなかった。一方で,テスト時
にオブジェクトを表す単語が用いられた場合,順序が同じ条件と同様に順序が異
なる条件でも学習効果が認められた。これらの結果は,オブジェクトの特徴情報
に基づいて時系列順序を抽出・学習し,さらに意味情報に基づいてオブジェクト
刺激をまとまりとして学習している可能性を示唆している。