Ossenkopp K P, Sorenson L, Mazmanian D S
Department of Psychology, University of Western Ontario, London, Ontario, N6A 5C2, Canada.
Behav Processes. 1994 Apr;31(2-3):129-44. doi: 10.1016/0376-6357(94)90001-9.
We examined the multivariate nature of open-field behavior in adult male rats (n = 26) by subjecting a longitudinal data set, obtained during 4 open-field test sessions (4 min in duration and spaced 48 h apart), to a three-way PARAFAC analysis. Unlike conventional two-way factor analytical models, the PARAFAC procedure allows for the direct factor analysis of 3-dimensional arrays, which then provided a unique factor solution to the longitudinal data set. The PARAFAC analysis extracted 2 factors: i) emotional reactivity and ii) exploratory behavior. These two factors changed in temporal prominence, with animals showing greater emotional reactivity on the first test session, and greater levels of exploration on the third and fourth test sessions. These results are in general agreement with previous findings which used more conventional factor analytic approaches. These findings indicate that multivariate procedures, such as the PARAFAC analysis, can be helpful in the quantitative characterization of behavioral phenomena in a more 'realistic' manner.
我们通过对在4次旷场试验（每次持续4分钟，间隔48小时）期间获得的纵向数据集进行三向平行因子分析（PARAFAC），研究了成年雄性大鼠（n = 26）旷场行为的多变量性质。与传统的双向因子分析模型不同，PARAFAC程序允许对三维阵列进行直接因子分析，从而为纵向数据集提供了独特的因子解。PARAFAC分析提取了2个因子：i）情绪反应性和ii）探索行为。这两个因子在时间上的突出程度有所变化，动物在第一次测试时表现出更高的情绪反应性，在第三次和第四次测试时表现出更高的探索水平。这些结果与之前使用更传统因子分析方法的研究结果总体一致。这些发现表明，诸如PARAFAC分析等多变量程序有助于以更“现实”的方式对行为现象进行定量表征。
