Закон Йеркса-Додсона: связь между стимулированием и успешностью научения

Abstract

The paper looks into the phenomenon of non-linearity of the link between success of learning and its stimulation (the Yerkes-Dodson law), arguing that the existing explanations are mostly based on the idea of overloading and disruption of appropriate psychological mechanisms when the task is too difficult or learning stimulation is too strong. The authors put forward an alternative explanation for non-linearity of the link between stimulation and success of learning: the Yerkes-Dodson effects may be caused by peculiarities of the learning algorithm employed by the brain. The authors argue that the above effects do not manifest themselves if learning is based on such instruction algorithms (widely used in non-declarative learning) as D. Hebb's reinforcement of a link by repetition, and outcome approximation described by the Rescorla-Wagner formula. An imitation experiment has demonstrated that the Yerkes-Dodson effects display themselves regularly if a person employs a more general variety of the above mentioned algorithms, i.e. the algorithm based on the scheme of neuron grids realizing the idea of non-linearity of the link between stimulation and outcome.