You are here

A closed-loop framework to investigate cerebellar learning in the decerebrated ferret preparation

TitleA closed-loop framework to investigate cerebellar learning in the decerebrated ferret preparation
Publication TypeMiscellaneous
Year of Publication2010
AuthorsZucca, R, Rasmussen, A, Herreros, I, Johansson, F, Verschure, PFMJ, Hesslow, G
Abstract

The cerebellum is a well-studied structure involved in motor learning and possibly in other cognitive tasks. Its synaptic organization and physiology is well defined and the analysis of cerebellar dependent forms of learning, such as Pavlovian eyelid conditioning, has revealed that climbing fibers (cf) from the Inferior Olive (conveying the unconditional stimulus information - US) can induce plastic changes in the cerebellar cortex. There is evidence that the inhibitory projection from the cerebellum to olive is a negative feedback pathway for controlling learning (Hesslow, 1986; Verschure et al., 2001). In order to investigate this feedback system, we have developed a technique where we can replace the cerebello-olivary control loop with a real-time physiologically constrained computer simulation. The simulation is directly interfaced through a recording and stimulation electrode to the in vivo preparation and allows us to control and manipulate the experimental protocol while tracking the same Purkinje cell over long periods of time. Here we present the properties of this closed loop framework for recording, stimulus generation and online analysis during eyeblink conditioning experiments. We have previously developed a decerebrated ferret preparation that allows stable and long (up to 18 h) single Purkinje cell recordings in vivo from a well-defined cerebellar microzone responsible for the control of the eyelid movement. Ferrets are trained using forelimb stimulation as conditioned stimulus (CS) and cf stimulation as US. Raw data are recorded from single Purkinje cells and spike events are detected online using an adaptive threshold crossing method. During each trial the activity of the cell in a time window that precedes the US presentation is monitored and stimulation of cfs is delivered if a certain criterion is satisfied. Random fluctuations in the cell firing during the time window under analysis and learning effects are tested online using a range of statistical analyses. In a pilot study we have used this technology to investigate how the closed loop stimulation of
cfs affects the temporal properties of the Purkinje cell response during continuous acquisition and extinction experiments. We observe that the Purkinje cell appears to follow a self-regulatory adjustment to the levels of cf activity. References: Hesslow, G. (1986) Inhibition of inferior olivary transmission by mesencephalic stimulation in the cat. Neurosci. Lett., 63, 76-80. Verschure, P. and Mintz, M. (2001) A real-time model of the cerebellar circuitry underlying classical conditioning: A combined simulation and robotics study. Neurocomputing, 38-40,1019-1024.

Full Text