ABSTRACT
DataAnalysis ....................................................................................................... 162 Extended-Return-Map ................................................................................... 162 Quanti‡cation of Differences between Two ERMs .................................... 163
Part 1: Experimental Study ................................................................................ 164 Method and Procedure .................................................................................. 164 Results .............................................................................................................. 165
Typical ERM Patterns of FI Responding ................................................ 165 ERM Patterns of Surrogate Data ............................................................. 166 Correlation Dimension of the Surrogate ERM Patterns ....................... 166 Development of the Temporal Dynamics of FI Responding ............... 167 Learning Curve of the FI Responding .................................................... 167
Part 2: Simulation Study .................................................................................... 171 Simulation Programs ..................................................................................... 171 Simulation Results.......................................................................................... 171
Lattice Structure ......................................................................................... 171 L-Type Structure ........................................................................................ 172 Triangle Structures .................................................................................... 174
Discussion ............................................................................................................ 177 Acknowledgment ................................................................................................ 178 Appendix 7.A Skinner-Box Experiments and FI Schedules ....................... 179 Appendix 7.B Simulation Programs .............................................................. 179 References ............................................................................................................. 181
The extended-return-map (ERM) was employed to analyze the interresponse time data of operant experiments using ‡xed-interval (FI) schedules and food reinforcement. After intensive training over numerous sessions, rats gradually developed several types of temporal patterns of lever-pressing (LP) behaviors, which were visualized through different patterns of data point distributions in the ERM. Analyses with randomly shuf°ed data sets con‡rmed that these patterns depended on the sequential order of the inter-response time data, indicating that they re°ected dynamics of the behavior. A procedure was developed to quantify the difference between patterns in the ERPs, thus enabling the comparison between sessions and animals. Simulations suggested that, in addition to the two-state break-and-burst responding, both multiple switches of behavioral states during the inter-reinforcement periods and the acceleration of LP rate should be taken into consideration for the dynamics found in the data.
