Behavioural Processes 57 (2002) 71 – 88 Habituation, memory and the brain: the dynamics of interval timing J.E.R. Staddon a, *, I.M Chelaru a , J.J. Higa b a Department of Psychology: Experimental, Duke Uniersity, Duraham, NC 27708, USA b Texas Christian Uniersity, Texas, USA Accepted 27 November 2001 Abstract Memory decay is rapid at first and slower later—a feature that accounts for Jost’s memory law: that old memories gain on newer ones with lapse of time. The rate-sensitive property of habituation — that recovery after spaced stimuli may be slower than after massed — provides a clue to the dynamics of memory decay. Rate-sensitive habituation can be modeled by a cascade of thresholded integrator units that have a counterpart in human brain areas identified by magnetic source imaging (MSI). The memory trace component of the multiple-time-scale model for habituation can provide a ‘clock’ that has the properties necessary to account for both static and dynamic properties of interval timing: static proportional and Weber-law timing as well as dynamic tracking of progressive, ‘impulse’ and periodic interval sequences. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Magnetic source imaging; Interval timing; Memory decay www.elsevier.com/locate/behavproc 1. Introduction When hungry animals are given a bit of food at regular intervals they soon learn to anticipate its arrival. When food delivery is response-indepen- dent, the procedure is termed temporal condition- ing; when it depends on the emission of an operant response after a fixed time, it is called a fixed-interval schedule. In the years since Pavlov and Skinner identified these procedures, numer- ous variants have been studied. All share two defining features: a to-be-timed interval, and a time marker, from which the interval is measured. In temporal conditioning, a fixed-interval (FI) schedule, or the response-initiated-delay (RID) schedule (Fig. 1, top), food delivery is the time marker. On procedures such as the peak-interval procedure or delay conditioning, some aspect of the transition from intertrial interval to trial onset constitutes the time marker. On spaced-respond- ing schedules, each response is a time marker. Usual practice in the study of interval timing is to expose the animal to the procedure for many intervals over many days, seeking the steady-state relation between the pattern of responding and the duration of the to-be-timed interval. Two properties have received a lot of attention. Pro- portional timing is the fact that many measures of timing are in the steady-state proportional to the to-be-timed interval: wait-time (time to first re- sponse, pause) on fixed-interval or RID schedules, * Corresponding author. E-mail address: staddon@psych.duke.edu (J.E.R. Staddon). 0376-6357/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S0376-6357(02)00006-2