Classical eyeblink conditioning with mixed interstimulus intervals: Temporal integration of response topography and neuronal correlates

• Main Author: Choi, June-Seek
• Language: ENG
• Published: ScholarWorks@UMass Amherst 1999
• Subjects: Neurology|Physiological psychology
• Online Access: https://scholarworks.umass.edu/dissertations/AAI9950143

The purpose of this thesis was to investigate the firing pattern of single neurons in cerebellar nucleus interpositus (NI) related to a conditioned response (CR) instilled by conditioning with mixed interstimulus intervals (ISIs). The conditioning with mixed ISIs is a protocol of classical conditioning that involves two different interstimulus intervals (ISIs) alternating randomly. The conditioned stimulus (CS) is a tone and the unconditioned stimulus (US) is an electrical shock to periorbital region that causes eyeblink. This protocol resulted in bimodal responses with peaks corresponding to the timing of the US. A related purpose of this investigation was to explore mechanisms of CR timing and temporal integration using a computational method. Sutton and Barto's (1990) temporal difference (TD) model with a complete serial compound (CSC) representation of CS was used to implement the idea of a neuron-like processing element that receives time-tagged inputs and adjusts their associative strength to generate an appropriate output adaptive to the given conditioning environment. The TD model with CSC representation of time can be aligned with the cerebellum. Within this context, there are several scenarios as to how time is segmented in the cerebellum and how this information is integrated to produce the CR. The current investigation presents evidence that single neurons of NI express a firing pattern closely related to the bimodal response. All forty-two CR-related units recorded in NI showed neuronal activity closely related to the time-course of eyeblink CRs, i.e. a neuronal activity pattern with two distinctive increases in firing rate. Most of the units preceded the behavioral response but the degree by which the neuronal activity preceded the behavioral response varied. Among the 42 CR-related units, 9 units responded to the tone CS with short latency(<100 ms), CS-locked activity. Among twenty-four units tested on a US-only trial, 22 units increased firing rate or remained at the same level, and 2 units decreased firing rate after US presentation. The CR topography on short-ISI reinforced trials was unimodal implying that the US has become a conditioned inhibitor. The corresponding neuronal activity of single neurons were also unimodal. The TD (CSC) model and its cerebellar implementation could account for the suppression of the second peak by employing a US-initiated timing cascade. The suppression of the neuronal activity on short-ISI trials suggests that the excitation and the inhibition could be expressed at the level of single neuron.