Feedback inhibition underlies new computational functions of cerebellar interneurons.

The function of a feedback inhibitory circuit between cerebellar Purkinje cells and molecular layer interneurons (MLIs) was defined by combining optogenetics, neuronal activity recordings both in cerebellar slices and in vivo, and computational modeling. Purkinje cells inhibit a subset of MLIs in the inner third of the molecular layer. This inhibition is non-reciprocal, short-range (less than 200 µm) and is based on convergence of one to two Purkinje cells onto MLIs. During learning-related eyelid movements in vivo, the activity of a subset of MLIs progressively increases as Purkinje cell activity decreases, with Purkinje cells usually leading the MLIs. Computer simulations indicate that these relationships are best explained by the feedback circuit from Purkinje cells to MLIs and that this feedback circuit plays a central role in making cerebellar learning efficient.

No neuron is an island: Homeostatic plasticity and over-constraint in a neural circuit.

To support computation the activity of neurons must vary within a useful range, which highlights one potential value of homeostatic plasticity. The interconnectedness of the brain, however, introduces the possibility that combinations of homeostatic mechanisms can produce over-constraint in which not all set points can be satisfied. We use a simulation of the cerebellum to investigate the potential for such conflict and its consequences. In this instance the conflict produces perpetual drift and eventual saturation of synaptic weights. We show that these problems can be resolved for this network by a particular combination of sites and rules for plasticity. We also show that simulations that implement these rules for homeostatic plasticity are more resistant to forgetting. These results illustrate the general principle that homeostatic plasticity within a system must not set up conflicts in which mutually exclusive set points exist and that one consequence can be perpetual induction of plasticity.

Medial Auditory Thalamus Is Necessary for Expression of Auditory Trace Eyelid Conditioning.

Transforming a brief sensory event into a persistent neural response represents a mechanism for linking temporally disparate stimuli together to support learning. The cerebellum requires this type of persistent input during trace conditioning to engage associative plasticity and acquire adaptively timed conditioned responses (CRs). An initial step toward identifying the sites and mechanisms generating and transmitting persistent signals to the cerebellum is to identify the input pathway. The medial auditory thalamic nuclei (MATN) are the necessary and sufficient source of auditory input to the cerebellum for delay conditioning in rodents and a possible input to forebrain sites generating persistent signals. Using pharmacological and computational approaches, we test (1) whether the necessity of MATN during auditory eyelid conditioning is conserved across species, (2) whether the MATN are necessary for the expression of trace eyelid CRs, and if so, (3)whether this relates to the generation of persistent signals. We find that contralateral inactivation of MATN with muscimol largely abolished trace and delay CRs in male rabbits. Residual CRs were decreased in amplitude, but CR timing was unaffected. Results from large-scale cerebellar simulations are consistent with previous experimental demonstrations that silencing only CS-duration inputs does not abolish trace CRs, and instead affects their timing. Together, these results suggest that the MATN are a necessary component of both the direct auditory stimulus pathway to the cerebellum and the pathway generating task-essential persistent signals.SIGNIFICANCE STATEMENT Persistent activity is required for working memory-dependent tasks, such as trace conditioning, and represents a mechanism by which sensory information can be used over time for learning and cognition. This neuronal response entails the transformation of a discrete sensory-evoked response into a signal that extends beyond the stimulus event. Understanding the generation and transmission of this stimulus transformation requires identifying the input sources necessary for task-essential persistent signals. We report that the medial auditory thalamic nuclei are required for the expression of auditory trace conditioning and suggest that these nuclei are a component of the pathway-generating persistent signals. Our study provides a foundation for testing circuit-level mechanisms underlying persistent activity in a cerebellar learning model with identified inputs and well defined behavioral outputs.

Cerebellar Processing Common to Delay and Trace Eyelid Conditioning.

Results from previous lesion studies have been interpreted as evidence that the cerebellar cortex plays different roles for delay and trace conditioning of eyelid responses. However, the cerebellar cortex is organized by parasagittal stripes of Purkinje cells (PCs) that converge onto common deep nucleus neurons and receive common or related climbing fiber inputs. Based on this organization, we hypothesized that cerebellar tasks involving the same response system, such as delay and trace eyelid conditioning, would engage the same PCs and that the relationships between PC activity and expression of behavioral responses would be similar for both tasks. To test these hypotheses, we used tetrode recordings from eyelid PCs in rabbits during expression of delay- and trace-conditioned eyelid responses. Previous recording studies during delay conditioning described a strong relationship between eyelid PC activity and the kinematics of conditioned eyelid responses. The present results replicate these findings for delay conditioning and show that the same relationship exists during trace eyelid conditioning. During transitions from delay to trace responding, the relationship between eyelid PCs and behavioral responses was relatively stable. We found that an inverse firing rate model tuned to predict PC activity during one training paradigm could then predict equally well the PC activity during the other training paradigm. These results provide strong evidence that cerebellar cortex processing is similar for delay and trace eyelid conditioning and that the parasagittal organization of the cerebellum, not the conditioning paradigm, dictate which neurons are engaged to produce adaptively timed conditioned responses.SIGNIFICANCE STATEMENT A variety of evidence from eyelid conditioning and other cerebellar-dependent behaviors indicates that the cerebellar cortex is necessary for learning and proper timing of cerebellar learned responses. Debates exist about whether trace eyelid conditioning data show that fundamentally different mechanisms operate in the cerebellum during tasks when input from the forebrain is necessary for learning. We show here that learning-related changes in a specific population of Purkinje cells control the timing and amplitude of cerebellar responses the same way regardless of the inputs necessary to learn the task. Our results indicate the parasagittal organization of the cerebellar cortex, not the complexity of inputs to the cerebellum, determines which neurons are engaged in the learning and execution of cerebellar-mediated responses.

A cerebellar adaptation to uncertain inputs.

Noise and variability are inherent and unavoidable features of neural processing. Despite this physiological challenge, brain systems function well, suggesting the existence of adaptations that cope with noise. We report a novel adaptation that the cerebellum implements to maintain correct responses in the face of ambiguous inputs. We found that under these conditions, the cerebellum used a probabilistic binary choice: Although the probability of behavioral response gradually increased or decreased depending on the degree of similarity between current and trained inputs, the size of response remained constant. That way the cerebellum kept responses adaptive to trained input corrupted by noise while minimizing false responses to novel stimuli. Recordings and analysis of Purkinje cells activity showed that the binary choice is made in the cerebellar cortex. Results from large-scale simulation suggest that internal feedback from cerebellar nucleus back to cerebellar cortex plays a critical role in implementation of binary choice.

Systematic variation of acquisition rate in delay eyelid conditioning.

Averaging artifacts inherent in group acquisition curves can mask behavioral phenomena that are potentially revealing in terms of underlying neural mechanisms. To address this, we implemented a behavioral analysis of 106 rabbits trained over 4 sessions using delay eyelid conditioning. Group results showed the typical monotonic increase in conditioned responses (CRs). For most subjects CRs first appeared (as indexed by the criterion of 8 CRs in 9 trials) during the first 18 trials of the second training session. Subdividing subjects according to the training block at which they met criterion revealed systematic differences in the subsequent rate that CR amplitudes increased, but not in asymptotic CR amplitudes. Subjects meeting criterion early in sessions showed more rapid increases in CR amplitude than those meeting criterion later in sessions. This effect was solely dependent on how early within a session criterion was met, as subjects meeting criterion at the beginning of the third and fourth sessions showed more rapid increases in CR amplitude than those meeting criterion after the first 18 trials of the second session. The exceptions were the 7% of the subjects that met criterion late in the first session. Their CR amplitudes increased at a rate similar to subjects meeting criterion early in sessions. These results suggest an interplay between consolidation processes and a previously reported short-term plasticity process that makes CR acquisition a nonmonotonic and complex function of the point during training sessions at which CRs first appear. (PsycINFO Database Record

Links Between Single-Trial Changes and Learning Rate in Eyelid Conditioning.

The discovery of single-trial learning effects, where the presence or absence (or the number) of climbing fiber inputs produces measureable changes in Purkinje cell response and in behavior, represents a major breakthrough in cerebellar learning. Among other things, these observations provide strong links between climbing fiber-mediated plasticity and cerebellar learning. They also demonstrate that cerebellar learning is stochastic, with each instantiation of a movement producing a small increment or decrement in gain. The sum of the small changes give rise to the macroscopic properties of cerebellar learning. We used a relatively large data set from another example of cerebellar-dependent learning, classical conditioning of eyelid responses, to attempt a behavioral replication and extension of single-trial learning effects. As a normal part of training, stimulus-alone trials provide instances where the climbing fiber response would be omitted, similar to non-climbing-fiber trials (gain down) during smooth pursuit training. The consequences of the stimulus-alone trial on the amplitude and timing of the conditioned response on the following paired trials were examined. We find that the amplitude of the conditioned response during the trial after a stimulus-alone trial (no climbing fiber input) was measurably smaller than the amplitude on the previous trials, and this single-trial effect on amplitude is larger for longer interstimulus intervals. The magnitude of the single-trial effect parallels the rate of extinction at different interstimulus intervals supporting the previously observed link between single-trial effects and learning.

Relating cerebellar purkinje cell activity to the timing and amplitude of conditioned eyelid responses.

How Purkinje cell (PC) activity may be altered by learning is central to theories of the cerebellum. Pavlovian eyelid conditioning, because of how directly it engages the cerebellum, has helped reveal many aspects of cerebellar learning and the underlying mechanisms. Theories of cerebellar learning assert that climbing fiber inputs control plasticity at synapses onto PCs, and thus PCs control the expression of learned responses. We tested this assertion by recording 184 eyelid PCs and 240 non-eyelid PCs during the expression of conditioned eyelid responses (CRs) in well trained rabbits. By contrasting the responses of eyelid and non-eyelid PCs and by contrasting the responses of eyelid PCs under conditions that produce differently timed CRs, we test the hypothesis that learning-related changes in eyelid PCs contribute to the learning and adaptive timing of the CRs. We used a variety of analyses to test the quantitative relationships between eyelid PC responses and the kinematic properties of the eyelid CRs. We find that the timing of eyelid PC responses varies systematically with the timing of the behavioral CRs and that there are differences in the magnitude of eyelid PC responses between larger-CR, smaller-CR, and non-CR trials. However, eyelid PC activity does not encode any single kinematic property of the behavioral CRs at a fixed time lag, nor does it linearly encode CR amplitude. Even so, the results are consistent with the hypothesis that learning-dependent changes in PC activity contribute to the adaptively timed expression of conditioned eyelid responses.

Cerebellar mechanisms of learning and plasticity revealed by delay eyelid conditioning.

The analysis of well-defined behaviors that require the cerebellum has helped reveal many key mechanisms operating in the cerebellum to mediate learning and feed-forward prediction. These systems include eyelid conditioning, adaptation of the vestibuloocular reflex, smooth pursuit eye movements, and arm-reaching tasks. This review focuses specifically on the variety of findings that have come from the use of eyelid conditioning to study the cerebellum. Phenomena discussed include sites and rules for plasticity, temporal coding and mechanisms of timing, cerebellar control of climbing fibers and its role in bidirectional learning, extinction of conditioned responses, and the phenomenon of savings.

Persistent activity in prefrontal cortex during trace eyelid conditioning: dissociating responses that reflect cerebellar output from those that do not.

Persistent neural activity, responses that outlast the stimuli that evoke them, plays an important role in neural computations and possibly in processes, such as working memory. Recent studies suggest that trace eyelid conditioning, which involves a temporal gap between the conditioned and unconditioned stimuli (the trace interval), requires persistent neural activity in a region of medial prefrontal cortex (mPFC). This persistent activity, which could be conveyed to cerebellum via a pathway through pons, may engage the cerebellum and allow for the expression of conditioned responses. Given the substantial reciprocity observed among many brain regions, it is essential to demonstrate that persistent responses in mPFC neurons are not simply a reflection of cerebellar feedback to the forebrain, leaving open the possibility that such responses could serve as input to the cerebellum. This concern is highlighted by studies showing that hippocampal learning-related activity is abolished by cerebellar inactivation. We inactivated the cerebellum while recording single-unit activity from the mPFC of rabbits trained with a forebrain-dependent trace eyelid conditioning procedure. We report that, whereas the responses of cells that show an onset of increased spike activity during the trace interval were abolished by cerebellar inactivation, persistent responses that begin during the conditioned stimulus and persisted into the trace interval were unaffected. Therefore, conditioned stimulus-evoked persistent responses remain the strongest candidate input pattern to support the cerebellar expression of learned responses.

Using a million cell simulation of the cerebellum: network scaling and task generality.

Several factors combine to make it feasible to build computer simulations of the cerebellum and to test them in biologically realistic ways. These simulations can be used to help understand the computational contributions of various cerebellar components, including the relevance of the enormous number of neurons in the granule cell layer. In previous work we have used a simulation containing 12000 granule cells to develop new predictions and to account for various aspects of eyelid conditioning, a form of motor learning mediated by the cerebellum. Here we demonstrate the feasibility of scaling up this simulation to over one million granule cells using parallel graphics processing unit (GPU) technology. We observe that this increase in number of granule cells requires only twice the execution time of the smaller simulation on the GPU. We demonstrate that this simulation, like its smaller predecessor, can emulate certain basic features of conditioned eyelid responses, with a slight improvement in performance in one measure. We also use this simulation to examine the generality of the computation properties that we have derived from studying eyelid conditioning. We demonstrate that this scaled up simulation can learn a high level of performance in a classic machine learning task, the cart-pole balancing task. These results suggest that this parallel GPU technology can be used to build very large-scale simulations whose connectivity ratios match those of the real cerebellum and that these simulations can be used guide future studies on cerebellar mediated tasks and on machine learning problems.

Persistent activity in a cortical-to-subcortical circuit: bridging the temporal gap in trace eyelid conditioning.

We have addressed the source and nature of the persistent neural activity that bridges the stimulus-free gap between the conditioned stimulus (CS) and unconditioned stimulus (US) during trace eyelid conditioning. Previous work has demonstrated that this persistent activity is necessary for trace eyelid conditioning: CS-elicited activity in mossy fiber inputs to the cerebellum does not extend into the stimulus-free trace interval, which precludes the cerebellar learning that mediates conditioned response expression. In behaving rabbits we used in vivo recordings from a region of medial prefrontal cortex (mPFC) that is necessary for trace eyelid conditioning to test the hypothesis that neurons there generate activity that persists beyond CS offset. These recordings revealed two patterns of activity during the trace interval that would enable cerebellar learning. Activity in some cells began during the tone CS and persisted to overlap with the US, whereas in other cells, activity began during the stimulus-free trace interval. Injection of anterograde tracers into this same region of mPFC revealed dense labeling in the pontine nuclei, where recordings also revealed tone-evoked persistent activity during trace conditioning. These data suggest a corticopontine pathway that provides an input to the cerebellum during trace conditioning trials that bridges the temporal gap between the CS and US to engage cerebellar learning. As such, trace eyelid conditioning represents a well-characterized and experimentally tractable system that can facilitate mechanistic analyses of cortical persistent activity and how it is used by downstream brain structures to influence behavior.

Multiple sites of extinction for a single learned response.

Most learned responses can be diminished by extinction, a process that can be engaged when a conditioned stimulus (CS) is presented but not reinforced. We present evidence that plasticity in at least two brain regions can mediate extinction of responses produced by trace eyelid conditioning, where the CS and the reinforcing stimulus are separated by a stimulus-free interval. We observed individual differences in the effects of blocking extinction mechanisms in the cerebellum, the structure that, along with several forebrain structures, mediates acquisition of trace eyelid responses; in some rabbits extinction was prevented, whereas in others it was largely unaffected. We also show that cerebellar mechanisms can mediate extinction when noncerebellar mechanisms are bypassed. Together, these observations indicate that trace eyelid responses can be extinguished via processes operating at more than one site, one in the cerebellum and one upstream in forebrain. The relative contributions of these sites may vary from animal to animal and situation to situation.

A subtraction mechanism of temporal coding in cerebellar cortex.

The temporally specific learning displayed by the cerebellum facilitates mechanistic analysis of neural timing and temporal coding. We report evidence for a subtraction-like mechanism of temporal coding in cerebellar cortex in which activity in a subset of granule cells specifically codes the interval between the offset of two mossy fiber inputs. In a large-scale cerebellar simulation, cessation of one of two ongoing mossy fiber inputs produces a robust temporal code in the population of granule cells. This activity supports simulation learning in response to temporal patterns of stimuli, even when those same stimuli do not support learning when presented individually. Using stimulation of mossy fiber inputs to the cerebellum as training stimuli in rabbits, we confirmed these unusual predictions in a cerebellum-dependent form of learning. Analysis of the simulations reveals a specific working hypothesis for this temporal subtraction process that involves interactions between granule cells and the inhibitory Golgi cells. The results suggest how feedforward inhibition, such as that present in the cerebellar cortex, can contribute to temporal coding.

A decrementing form of plasticity apparent in cerebellar learning.

Long-term synaptic plasticity is believed to underlie the capacity for learning and memory. In the cerebellum, for example, long-term plasticity contributes to eyelid conditioning and to learning in eye movement systems. We report evidence for a decrementing form of cerebellar plasticity as revealed by the behavioral properties of eyelid conditioning in the rabbit. We find that conditioned eyelid responses exhibit within-session changes that recover by the next day. These changes, which increase with the interstimulus interval, involve decreases in conditioned response magnitude and likelihood as well as increases in latency to onset. Within-subject comparisons show that these changes differ in magnitude depending on the type of training, arguing against motor fatigue or changes in motor pathways downstream of the cerebellum. These phenomena are also observed when stimulation of mossy fibers substitutes for the conditioned stimulus, suggesting that changes take place within the cerebellum or in downstream efferent pathways. Together, these observations suggest a plasticity mechanism in the cerebellum that is induced during training sessions and fades within 23 h. To formalize this hypothesis more specifically, we show that incorporating a short-lasting potentiation at the granule cell to Purkinje cell synapses in a computer simulation of the cerebellum reproduces these behavioral effects. We propose the working hypothesis that the presynaptic form of long-term potentiation observed at these synapses is reversed by time rather than by a corresponding long-term depression. These results demonstrate the utility of eyelid conditioning as a means to identify and characterize the rules that govern input to output transformations in the cerebellum.

Eyelid conditioning to a target amplitude: adding how much to whether and when.

Conceptual and practical advantages of pavlovian eyelid conditioning facilitate analysis of cerebellar computation and learning. Even so, eyelid conditioning procedures are unrealistic in an important way. The error signal to the olivocerebellar system does not decrease as learning adapts response amplitude or gain. This inherently limits the utility of eyelid conditioning for studies investigating how cerebellar learning mechanisms acquire and store an adaptive response amplitude. We report the development and characterization of a training procedure in which conditioned response amplitude is brought under experimental control with contingencies that more closely parallel natural conditions. In this procedure, the delivery of the unconditioned stimulus (US) is made contingent on conditioned response amplitude: the US is delivered for responses that fail to reach a specified target amplitude and is omitted for responses that meet or exceed the target. We find that rabbits trained with either a tone or with mossy fiber stimulation as the conditioned stimulus learn responses that approach target amplitudes ranging from 2 to 5 mm. Inactivating the interpositus nucleus with muscimol infusions abolished these conditioned responses, indicating that cerebellar involvement in eyelid conditioning is not tied explicitly to the use of pavlovian procedures. Together with previous studies, these data suggest that response amplitude is learned and encoded in the cerebellum during eyelid conditioning. As such, these results provide a foundation for systematic and controlled investigations of the cerebellar mechanisms that learn and encode the proper amplitude of adaptive movements.

Temporal patterns of inputs to cerebellum necessary and sufficient for trace eyelid conditioning.

Trace eyelid conditioning is a form of associative learning that requires several forebrain structures and cerebellum. Previous work suggests that at least two conditioned stimulus (CS)-driven signals are available to the cerebellum via mossy fiber inputs during trace conditioning: one driven by and terminating with the tone and a second driven by medial prefrontal cortex (mPFC) that persists through the stimulus-free trace interval to overlap in time with the unconditioned stimulus (US). We used electric stimulation of mossy fibers to determine whether this pattern of dual inputs is necessary and sufficient for cerebellar learning to express normal trace eyelid responses. We find that presenting the cerebellum with one input that mimics persistent activity observed in mPFC and the lateral pontine nuclei during trace eyelid conditioning and another that mimics tone-elicited mossy fiber activity is sufficient to produce responses whose properties quantitatively match trace eyelid responses using a tone. Probe trials with each input delivered separately provide evidence that the cerebellum learns to respond to the mPFC-like input (that overlaps with the US) and learns to suppress responding to the tone-like input (that does not). This contributes to precisely timed responses and the well-documented influence of tone offset on the timing of trace responses. Computer simulations suggest that the underlying cerebellar mechanisms involve activation of different subsets of granule cells during the tone and during the stimulus-free trace interval. These results indicate that tone-driven and mPFC-like inputs are necessary and sufficient for the cerebellum to learn well-timed trace conditioned responses.

Cerebellar cortex contributions to the expression and timing of conditioned eyelid responses.

We used micro-infusions during eyelid conditioning in rabbits to investigate the relative contributions of cerebellar cortex and the underlying deep nuclei (DCN) to the expression of cerebellar learning. These tests were conducted using two forms of cerebellum-dependent eyelid conditioning for which the relative roles of cerebellar cortex and DCN are controversial: delay conditioning, which is largely unaffected by forebrain lesions, and trace conditioning, which involves interactions between forebrain and cerebellum. For rabbits trained with delay conditioning, silencing cerebellar cortex by micro-infusions of the local anesthetic lidocaine unmasked stereotyped short-latency responses. This was also the case after extinction as observed previously with reversible blockade of cerebellar cortex output. Conversely, increasing cerebellar cortex activity by micro-infusions of the GABA(A) antagonist picrotoxin reversibly abolished conditioned responses. Effective cannula placements were clustered around the primary fissure and deeper in lobules hemispheric lobule IV (HIV) and hemispheric lobule V (HV) of anterior lobe. In well-trained trace conditioned rabbits, silencing this same area of cerebellar cortex or reversibly blocking cerebellar cortex output also unmasked short-latency responses. Because Purkinje cells are the sole output of cerebellar cortex, these results provide evidence that the expression of well-timed conditioned responses requires a well-timed decrease in the activity of Purkinje cells in anterior lobe. The parallels between results from delay and trace conditioning suggest similar contributions of plasticity in cerebellar cortex and DCN in both instances.

Interactions between prefrontal cortex and cerebellum revealed by trace eyelid conditioning.

Eyelid conditioning has proven useful for analysis of learning and computation in the cerebellum. Two variants, delay and trace conditioning, differ only by the relative timing of the training stimuli. Despite the subtlety of this difference, trace eyelid conditioning is prevented by lesions of the cerebellum, hippocampus, or medial prefrontal cortex (mPFC), whereas delay eyelid conditioning is prevented by cerebellar lesions and is largely unaffected by forebrain lesions. Here we test whether these lesion results can be explained by two assertions: (1) Cerebellar learning requires temporal overlap between the mossy fiber inputs activated by the tone conditioned stimulus (CS) and the climbing fiber inputs activated by the reinforcing unconditioned stimulus (US), and therefore (2) trace conditioning requires activity that outlasts the presentation of the CS in a subset of mossy fibers separate from those activated directly by the CS. By use of electrical stimulation of mossy fibers as a CS, we show that cerebellar learning during trace eyelid conditioning requires an input that persists during the stimulus-free trace interval. By use of reversible inactivation experiments, we provide evidence that this input arises from the mPFC and arrives at the cerebellum via a previously unidentified site in the pontine nuclei. In light of previous PFC recordings in various species, we suggest that trace eyelid conditioning involves an interaction between the persistent activity of delay cells in mPFC-a putative mechanism of working memory-and motor learning in the cerebellum.

Synapses, circuits, and the ontogeny of learning.

This article summarizes the proceedings of a symposium organized by Mark Stanton and Pamela Hunt and presented at the annual meeting of the International Society for Developmental Psychobiology. The purpose of the symposium was to review recent advances in neurobiological and developmental studies of fear and eyeblink conditioning with the hope of discovering how neural circuitry might inform the ontogenetic analyses of learning and memory, and vice versa. The presentations were: (1) Multiple Brain Regions Contribute to the Acquisition of Pavlovian Fear by Michael S. Fanselow; (2) Expression of Learned Fear: Appropriate to Age of Training or Age of Testing by Rick Richardson; (3) Trying to Understand the Cerebellum Well Enough to Build One by Michael D. Mauk; and (4) The Ontogeny of Eyeblink Conditioning: Neural Mechanisms by John H. Freeman. Taken together, these presentations converge on the conclusions that (1) seemingly simple forms of associative learning are governed by multiple "engrams" and by temporally dynamic interactions among these engrams and other circuit elements and (2) developmental changes in these interactions determine when and how learning emerges during ontogeny.