Bistability of mixed states in a neural network storing hierarchical patterns

We discuss the properties of equilibrium states in an autoassociative memor y model storing hierarchically correlated patterns thereafter, hierarchical patterns). We will show that symmetric mixed states thereafter, mixed stat es) are bistable on the associative memory model storing the hierarchical p atterns in a region of the ferromagnetic phase. This means that the first-o rder transition occurs in this ferromagnetic phase. We treat these contents with a statistical mechanical method (SCSNA) and by computer simulation. F inally, we discuss a physiological implication of this model. Sugase er al (1999 Nature 400 869) analysed the time-course of the information carried b y the firing of face-responsive neurons in the inferior temporal cortex. We also discuss the relation between the theoretical results and the physiolo gical experiments of Sugase et al.