Cerebellum and Learning: A Complex Problem

Annals of the New York Academy of Sciences, 1991

Kehoe's 'serial compound' experiments demonstrate conditioned reflex [CR] peaks leading or lagging the unconditioned stimulus [US], a phenomenon which the 'spectrum of microstimuli' hypotheses cannot explain. ‘Serial compound’ refers to a conditioning paradigm during which two different conditioned stimuli [CSs] are presented in succession (hence with different interstimulus intervals) BEFORE the US is administered. The ‘spectrum of microstimuli hypothesis’ – in the cerebellar context - assumes that the CS-onset signal triggers a timing mechanism [TM] that is based on interactions in the granular layer network : a ‘network based TM’ [NBTM]. Classically it is believed that a consistent US induces LTD at the parallel fiber-Purkinje cell [PuC] synapses of the granule cells which happen to be recurrently activated by the NBTM around US time, ultimately causing the PuCs to pause around US time, disinhibiting the nucleus interpositus [NI] to produce a CR. It will be shown how a TM that is triggered twice (which in turn is an effect of ‘cross-modal transfer’ [CMT] ) can lead to displaced CR peaks. The predicted displacements do not concur completely with those observed by Kehoe, but a tentative explanation for this discrepancy will be offered. However, it will be argued that an NBTM which lends itself to be triggered twice, is inevitably also an NBTM which exhibits CMT in all testing situations – not only in Kehoe’s experimental setup, and this argument, amongst others, severely questions the concept that the cerebellum engages an NBTM for nictitating membrane response and similar conditioning purposes. An intracellular TM in the PuCs as demonstrated by Johansson et al. is a much better candidate to ‘implement’ this kind of conditioning and to explain Kehoe’s observations, since it can escape CMT in other situations thanks to a (hypothesized) CS segregating mechanism in the NI. ‘Displaced’ CR peaks by themselves may also explain why ‘Kamin blocking’ fails when serial compound training follows full training with one of the CSs, while it is observed to occur when simultaneous compound training follows.

Proceedings of The National Academy of Sciences, 2002

The cerebellum is considered a brain structure in which memories for learned motor responses (e.g., conditioned eyeblink responses) are stored. Within the cerebellum, however, the relative importance of the cortex and the deep nuclei in motor learning͞memory is not entirely clear. In this study, we show that the cerebellar cortex exerts both basal and stimulus-activated inhibition to the deep nuclei. Sequential application of a ␥-aminobutyric acid type A receptor (GABAAR) agonist and a noncompetitive GABAAR antagonist allows selective blockade of stimulus-activated inhibition. By using the same sequential agonist and antagonist methods in behaving animals, we demonstrate that the conditioned response (CR) expression and timing are completely dissociable and involve different inhibitory inputs; although the basal inhibition modulates CR expression, the conditioned stimulus-activated inhibition is required for the proper timing of the CR. In addition, complete blockade of cerebellar deep nuclear GABAARs prevents CR acquisition. Together, these results suggest that different aspects of the memories for eyeblink CRs are encoded in the cerebellar cortex and the cerebellar deep nuclei.

Behavioral and Neural Biology, 1981

Male Long-Evans rats chronically implanted with an electrode in the mesencephalic reticular formation were tested in a latent inhibition paradigm to contrast the idea that, under our conditions, stimulation of the mesencephalic reticular formation enhances memory processing with the notion that such stimulation is aversive. In the preexposure phase of the experiment some animals were given experience with the to-be-conditioned stimulus, a tone. Others received the tone followed by stimulation of the mesencephalic reticular formation or a flashing light. The animals were then fear conditioned by pairing the tone with footshock. In the test, drink latencies in the presence of the tone indicated that animals given preexposure to the tone followed by stimulation of the mesencephalic reticular formation or the flashing light showed increased latent inhibition relative to controls given preexposure to the tone but no stimulation. These results strongly suggest that stimulation of the mesencephalic reticular formation is not aversive but, rather, acts to enhance memory processing. The idea that the flashing light may also serve to enhance memory processing is discussed.

Ten male albino rabbits were implanted with stimulating electrodes in the lateral reticular nucleus (LRN). These rabbits were given paired classical conditioning training of the nictitating membrane response with stimulation of the LRN as the conditioned stimulus (CS). Each rabbit was given daily training sessions until it consistently made conditioned responses (CRs). Each rabbit then received an aspiration lesion of cerebellar cortex: the ipsilateral anso-paramedian lobule (n = 6), the anterior or central vermis (n = 2), the central vermis and ansiform lobule (n = 1), or the central vermis and paramedian lobule (n = 1). After recovery, these rabbits were again given paired classical conditioning training with LRN stimulation as the CS. The rabbits with anso-paramedian lesions did not retain the CR after the lesion, but were able to relearn it. The rabbits with lesions of the vermis, the vermis and ansiform, or the vermis and paramedlan retained the CR after the lesion. These results are contrasted with previous results, which show that after aspiration of the anso-paramedian lobule, the conditioned response is not retained or relearned when stimulation of the dorsolateral pontine nucleus (DLPN) is used as a CS. The differences between the mossy fiber outputs of the LRN and DLPN may account for this discrepancy. Different regions of the cerebellum are apparently involved in retention of classically conditioned responses depending on the population of mossy fibers carrying the CS information.

NeuroReport, 2007

Associative learning in the cerebellum has previously focused on single movements. In eyeblink conditioning, for instance, a subject learns to blink at the right time in response to a conditional stimulus (CS), such as a tone that is repeatedly followed by an unconditional corneal stimulus (US). During conditioning, the CS and US are transmitted by mossy/parallel fibers and climbing fibers to cerebellar Purkinje cells that acquire a precisely timed pause response that drives the overt blink response. The timing of this conditional Purkinje cell response is determined by the CS–US interval and is independent of temporal patterns in the input signal. In addition to single movements, the cerebellum is also believed to be important for learning complex motor programs that require multiple precisely timed muscle contractions, such as, for example, playing the piano. In the present work, we studied Pur-kinje cells in decerebrate ferrets that were conditioned using electrical stimulation of mossy fiber and climbing fiber afferents as CS and US, while alternating between short and long interstimulus intervals. We found that Purkinje cells can learn double pause responses , separated by an intermediate excitation, where each pause corresponds to one interstimulus interval. The results show that individual cells can not only learn to time a single response but that they also learn an accurately timed sequential response pattern. cerebellum | Purkinje cells | learning | timing | classical conditioning P laying the piano, typing on your keyboard, and uttering a sentence are all typical examples of complex behaviors. Although they involve simple movements as building blocks, they need to be executed with great temporal precision and in a specific sequential order. Utter phonemes in the wrong order or with incorrect timing, and the result is incomprehensible. Learning of accurately timed movements is exemplified by classical conditioning of motor responses such as the eyeblink response (1–3). In the simplest case, a neutral conditional stimulus (CS), usually a tone, is followed by a blink-eliciting unconditional stimulus (US), for example, a puff of air to the cornea, after a fixed interval (the interstimulus interval, ISI). The subject learns to emit a conditional blink response (CR) that is timed so that the maximum amplitude is reached close to the time of US onset. Eyeblink conditioning depends on the cerebellar cortex (4–6) and the overt CRs are driven by learned pause responses in the spontaneously active cerebellar Purkinje cells (7–9). These cells receive the CS signal via the mossy fiber/parallel fiber system, and the US signal via the climbing fibers from the inferior olive (10, 11). The conditional Purkinje cell response (Purkinje cell CR, or PcCR) is elicited by input from the parallel fibers (12), triggering a delayed and adaptively timed pause in the Purkinje cell's simple spike firing (13), illustrated in Fig. 1A. Importantly, the PcCR is accurately timed even when the CS consists of a uniform and repetitive train of electrical pulses applied directly to the mossy fiber or parallel fiber afferents; that is, when the CS input signal to the cell contains no temporal code (14). Eyeblink conditioning with mixed ISIs, that is, using different intervals between CS and US on alternating trials, produces more complex temporal patterns in the blink CRs. The responses are less stereotypical than those obtained with a single ISI, often consisting of long-duration blinks or multipeaked blinks that form response sequences (see Fig. 1B for illustrations), with temporal profiles adapted to the ISIs (15, 16). Excitatory response patterns that match double-peaked blink responses have also been observed in the anterior interpositus nucleus, the downstream target of the blink controlling areas in the cerebellar cortex (17). Several important questions are raised by these findings. First, it may be asked whether a Purkinje cell can learn more than one interval; that is, can the cell learn to respond to a uniform repetitive CS containing no temporal code, with sequences corresponding to the long-duration or double-peaked eyeblink CRs, after training with alternating ISIs? There are data that suggest this is the case, but they consist only of unsystematic observations in a very small number of subjects: three decerebrate ferrets (12, 13) and two intact rabbits (18). Second, if a cell can learn a response sequence, the question may be asked of how the components of such a sequence are related to each other. For instance, is a double pause response a composite of separate response components, or should the whole response sequence be regarded as a single unit? If they are composites of separate components, then it should be possible to learn each component independently. Conversely, if it is a single response unit, then the whole sequence conceivably also could be elicited with shorter versions of the CS, as has previously been shown for single PcCRs (19) and eyeblink CRs (20). To answer these questions, we studied Purkinje cells in de-cerebrate ferrets during training with a classical conditioning paradigm, using a mossy fiber CS and climbing fiber US, with alternating ISIs. Results We made extracellular recordings in vivo, lasting for 3–12 h, of activity in 27 Purkinje cells in 22 decerebrated and immobilized male ferrets. Behavioral data were thus not collected, as the long extracellular recording sessions required immobilization for sufficient tissue stability. All recordings were made in a micro-zone within the cerebellar C3 zone that controls the conditional blink response (7). The CS was a uniform and repetitive stimulus consisting of direct electrical stimulation of mossy fibers at 50 Hz for 600 ms (or 800 ms in two cases). Climbing fibers were Significance Learning is thought to rely on the strengthening or weakening of synapses. However, we have previously shown that a neu-ron can also learn when to time its response so that the timing reflects the interval between two stimuli, without any temporal information in the input signal. Here, we report that a neuron can even learn a sequence of at least two, and probably more, accurately timed responses. A single cell is in a sense " programmable, " and can encode a temporal response pattern. This means that the nature of what a cell can learn is very different from the traditional view, and that the information storage capacity may be far greater.

Learning and Motivation, 1999

Historically, there has been little consensus regarding the outcome of backward conditioning trials in Pavlovian conditioning. One view states that few US→CS trials will produce an excitatory association, but with additional trials the excitatory association will wane and eventually give way to an inhibitory association ). An alternative view is that the initial association remains intact over further training trials, but the subject additionally learns the backward temporal relationship between the CS and the US (Barnet & Miller, 1996). Toward testing these views, we conducted four parametric experiments using conditioned suppression by rats to examine the development of excitatory and inhibitory response potentials as a function of the number of trials. In all experiments, animals received a low (4), moderate (16), or high (96) number of backward conditioning training. In Experiments 1 and 2, conditioned inhibition was assessed with summation and retardation tests, respectively, and more inhibition was found with more backward pairings. In Experiment 3, first-order excitatory responding was observed only with low levels of training. In Experiment 4, robust second-order excitatory responding was seen following low and high levels of US→CS training. The results are discussed in terms of Heth's views and the temporal coding hypothesis, a recent model of Pavlovian conditioned responding.

Integrative physiological and behavioral science : the official journal of the Pavlovian Society

A large body of evidence indicates that the cerebellum is essential for the acquisition, retention, and expression of the standard delay conditioned eyeblink response and that the basic memory trace appears to be established in the anterior interpositus nucleus (IP). Adaptive timing of the conditioned response (CR) is a prominent feature of classical conditioning-the CR peaks at the time of onset of the unconditioned stimulus (US) over a wide range of CS-US interstimulus intervals (ISI). A key issue is whether this timing is established by the cerebellar circuitry or prior to the cerebellum. In this study timing of conditioned eyeblink responses established via electrical stimulation of the interpositus nucleus as a conditioned stimulus (CS) was analyzed prior to and following modification of the CS-US interval in well-trained rabbits. Consistent with previous results, learning under these conditions is very rapid and robust. The CR peak eyeblink latencies are initially timed to the...

1998

Intradendritic recordings in Purkinje cells from a defined area in parasaggital slices of cerebellar lobule HVI, obtained after rabbits were given either paired (classical conditioning) or explicitly unpaired (control) presentations of tone and periorbital electrical stimulation, were used to assess the nature and duration of conditioning-specific changes in Purkinje cell dendritic membrane excitability. We found a strong relationship between the level of conditioning and Purkinje cell dendritic membrane excitability after initial acquisition of the conditioned response. Moreover, conditioning-specific increases in Purkinje cell excitability were still present 1 month after classical conditioning. Although dendritically recorded membrane potential, input resistance, and amplitude of somatic and dendritic spikes were not different in cells from paired or control animals, the size of a potassium channel-mediated transient hyperpolarization was significantly smaller in cells from animals that received classical conditioning. In slices of lobule HVI obtained from naive rabbits, the conditioning-related increases in membrane excitability could be mimicked by application of potassium channel antagonist tetraethylammonium chloride, iberiotoxin, or 4-aminopyridine. However, only 4-aminopyridine was able to reduce the transient hyperpolarization. The pharmacological data suggest a role for potassium channels and, possibly, channels mediating an I A -like current, in learning-specific changes in membrane excitability. The conditioning-specific increase in Purkinje cell dendritic excitability produces an afterhyperpolarization, which is hypothesized to release the cerebellar deep nuclei from inhibition, allowing conditioned responses to be elicited via the red nucleus and accessory abducens motorneurons.

Electroencephalography and Clinical Neurophysiology, 1967

Neural Networks, 2013

According to a widely held assumption, the main mechanism underlying motor learning in the cerebellum, such as eyeblink conditioning, is long-term depression (LTD) of parallel fibre to Purkinje cell synapses. Here we review some recent physiological evidence from Purkinje cell recordings during conditioning with implications for models of conditioning. We argue that these data pose four major challenges to the LTD hypothesis of conditioning. (i) LTD cannot account for the pause in Purkinje cell firing that is believed to drive the conditioned blink. (ii) The temporal conditions conducive to LTD do not match those for eyeblink conditioning. (iii) LTD cannot readily account for the adaptive timing of the conditioned response. (iv) The data suggest that parallel fibre to Purkinje cell synapses are not depressed after learning a Purkinje cell CR. Models based on metabotropic glutamate receptors are also discussed and found to be incompatible with the recording data.

The Journal of Physiology, 1996

Frontiers in Computational Neuroscience, 2010

type of input, a "teaching" signal carried by climbing fibers. A simple yet general implementation of this conceptual framework is the adaptive filter, a generic signal-processing device that has wide utility in control theory and other applications (Widrow, 1985), and was first formally applied to the cerebellum by Fujita (1982). Because the cerebellum has such a homogeneous structure, it can be hypothesized that each microcircuit implements the same basic algorithm and that a particular behavioral task is specified by the information content of its climbing fiber input and by the target of its output. Adaptive filter models have been used successfully for many of the cerebellum-dependent tasks referred to above (references given in Dean et al., 2010). Classical conditioning of eyeblink and skeletal muscle reflexes has been shown to be dependent on the cerebellum (see review by Yeo and Hesslow, 1998) and experimental analysis of the neural circuitry underlying eyeblink conditioning suggests mechanisms that fit well within the Marr-Albus framework. It appears that the conditioned stimulus (CS) is a "context" delivered to the cerebellum by the mossy fiber system, the unconditioned stimulus (US) is a "teaching" signal borne by the climbing fiber input from the inferior olive, and the resultant cerebellar output drives the conditioned response (CR) as a motor output (Hesslow and Yeo, 2002;

Animal Learning & Behavior, 1994