Cerebellar climbing fibers encode expected reward size

Larry, Noga; Yarkoni, Merav; Lixenberg, Adi; Joshua, Mati

doi:10.1101/533653

Cited by 29 publications

(58 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although suggestive, our results in this study about the CS signals providing a 333 reward expectation signal during learning combined with the results of our previous study 13 about 334 the SS signals providing a reinforcement learning contingent error signal fits well in the framework 335 that the cerebellum might be working in tandem with the basal ganglia in effectively driving 336 reward-related behavior. 337 Furthermore, we did not find any relationship between the CS activity and the SS 338 activity 7,25 . One might have expected that a CS signal could have served as a teaching signal for 339 the delta epoch of SS during learning if the classical error correcting framework were to apply.…”

Section: Discussion: 267mentioning

confidence: 60%

“…This was not at all the case (Fig 6). There are several reasons why complex spike signals are 341 unlikely to play the role of a teaching signal in our experiment 20 . First, it is unlikely that the 342 difference in the simple spike rate of more than 30 sp/s in the delta epochs between consecutive 343 correct and wrong trials could be caused solely by synaptic depression elicited by complex spikes 344 which has only been shown to cause a maximum of 8-10 sp/s changes in SS activity (with the 345 longest CS waveforms) 2,5 .…”

Section: Discussion: 267mentioning

confidence: 96%

“…In contrast, in the oculomotor system, for smooth pursuit 16 and the 278 otololithic vesitibulo-ocular reflex 17 , CS activity describes the movement itself, with CS inhibition 279 of simple spikes providing a reciprocal modulation in the simple spike activity. In well-learned 280 reaching tasks in the monkey, the CS of different P-cells report kinematic errors in position, 281 velocity, and acceleration 18 282 There is recent overwhelming evidence that the cerebellum is implicated in tasks beyond 283 simple motor learning 8,10,13,19,20 . McCormick et al 21 originally showed that the cerebellum is 284 necessary for the Pavolvian association task of eye-blink conditioning, and since then, several 285 others have shown that in eye-blink conditioning 6 and in other forms of simple classical 286 conditioning 8,9 in the mouse, CS activity provides a temporal-difference prediction error.…”

Section: Discussion: 267mentioning

confidence: 99%

See 2 more Smart Citations

Complex spikes encode reward expectation signals during visuomotor association learning

Sendhilnathan

Ipata

Me³

2019

Preprint

View full text Add to dashboard Cite

1Climbing fiber input to Purkinje cells has been thought to instruct learning related changes 2 in simple spikes and cause behavioral changes through an error-based learning mechanism. 3 Although, this framework explains simple motor learning, it cannot be extended to learning higher-4 order skills. Recently the cerebellum has been implicated in a variety of cognitive tasks and 5 reward-based learning. Here we show that when a monkey learns a new visuomotor association, 6 complex spikes predict the time of the beginning of the trial in a learning independent manner as 7 well as encode a learning contingent reward expectation signal after the stimulus onset and reward 8 delivery. These complex spike signals are unrelated to and were unlikely to instruct the reward 9 based signal found in the simple spikes. Our results provide a more general role of complex spikes 10 in learning and higher-order processing while gathering evidence for their participation in reward 11 based learning. 12 13 Introduction: 14The cerebellum has been classically considered to be one of the centers for supervised 15 learning in the brain, where the predicted results of movement are compared with the animal's 16 actual performance in relation to their sensory experience, in order to correct the errors in the 17 action that led to the mismatch. The cerebellar cortex has been posited to achieve this via its two 18 distinct types of inputs to its principle output cells, the Purkinje cells (P-cells). First, an efference 19 copy of the ongoing plan generated by other areas in the brain, communicated to the P-cells 20 through mossy fibers, and read out as high frequency simple spikes (SS); and second, a putative 21 instruction signal, motioning unexpected events, communicated through the projections from the 22 inferior olive, the climbing fibers, which evoke low frequency complex spikes (CS). The precisely 23 timed relationship between the coincidence of complex spikes and simple spikes has been to shown 24 to cause a long-term depression at the granule cell->P-cell synapse thereby supervising the 25 information being learned at the level of P-cells 1 . 26The CS have several characteristics: First, they fire with extremely low firing rates 27 (baseline: 0.5-1 Hz, to ~10 Hz), second, they encode sparsely by only firing in about 20-30% of 28 the trials 2,3 , third: some are predominantly evoked by mismatches in sensory predictions 4 , fourth, 29 they can encode information in their rate of firing 2 , timing of firing 3 or the duration of spikes 5 , 30fifth, presence of a single complex on a trial could potentially drive learning and might cause about 31 1 -10 sp/s change in the magnitude of simple spike firing rate in the next trial 3,5 . However, due to 32 the combined effect of low frequency, sparseness and low amount changes that it causes to simple 33 spike, significant changes in behavior, as a consequence of learning, only happens over several 34 tens of trials. This flow of information and circuitry explains many simple motor learning 35 b...

show abstract

Section: Discussion: 267mentioning

confidence: 60%

Section: Discussion: 267mentioning

confidence: 96%

Section: Discussion: 267mentioning

confidence: 99%

See 1 more Smart Citation

Complex spikes encode reward expectation signals during visuomotor association learning

Sendhilnathan

Ipata

Me³

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…Such interactions are likely to be supported by 299 the recent demonstration of direct anatomical connections between the cerebellum and striatum 300 (67). These bidirectional connections could explain recent neural findings from rodents showing 301 that the cerebellum, besides processing direction-related errors, also represents various aspects 302 of reward-related information during task performance (68)(69)(70)(71). Together, this emerging evidence 303 suggests that error-based and reward-based learning processes are closely intertwined at both 304 the behavioral and neural levels.…”

mentioning

confidence: 87%

Variability in error-based and reward-based human motor learning is associated with entorhinal volume

Brouwer

Rashid

Flanagan

et al. 2020

Preprint

View full text Add to dashboard Cite

Variation in error-based and reward-based human motor1 learning is related and associated with entorhinal volume 2 3 Anouk J. de Brouwer1, Mohammad R. Rashid2, J. Abstract 13Error-based and reward-based processes are critical for motor learning, and are thought to be 14 supported via distinct neural pathways. However, recent behavioral work in humans suggests that 15 both learning processes are supported by cognitive strategies and that these contribute to 16 individual differences in motor learning ability. While it has been speculated that medial temporal 17 lobe regions may support this strategic component to learning, direct evidence is lacking. Here 18 we first show that faster and more complete learning during error-based visuomotor adaptation is 19 associated with better learning during reward-based shaping of reaching movements. This result 20 suggests that strategic processes, linked to faster and better learning, drive individual differences 21 in both error-based and reward-based motor learning. We then show that right entorhinal cortex 22 volume was larger in good learning individuals-classified across both motor learning tasks-23 compared to their poorer learning counterparts. This suggests that strategic processes underlying 24 both error-and reward-based learning are linked to neuroanatomical differences in entorhinal 25 cortex. 26 27 Significance Statement 30While it is widely appreciated that humans vary greatly in their motor learning abilities, little is 31 known about the processes and neuroanatomical bases that underlie these differences. Here, 32 using a data-driven approach, we show that individual variability in error-based and reward-based 33 motor learning is tightly linked, and related to the use of cognitive strategies. We further show that 34 structural differences in entorhinal cortex predict this intersubject variability in motor learning, with 35 larger entorhinal volumes being associated with better overall error-based and reward-based 36 learning. Together, these findings provide support for the notion that the ability to recruit strategic 37 processes underlies intersubject variability in both error-based and reward-based learning, which 38 itself may be linked to structural differences in medial temporal cortex. 39When considering the neural bases of such interactions, it is important to distinguish between two 62 main forms of learning. Error-based learning is the form of learning by which we refine and adjust 63 our movements to changes in the body or the environment based on observable errors, such as 64 when missing the bullseye in archery. Such learning is thought arise from two separate processes 65 acting in parallel: An implicit process that is driven by the error between predicted and observed 66 sensory feedback (17,18) and an explicit process that is driven by the error between the target 67 and the sensory feedback (i.e., the task error) (19). The implicit process is nonconscious (i.e., 68 resistant to voluntary control), adapts and de-adapts gradually and is relian...

show abstract

“…Simple spike plasticity, under control of climbing fiber activity, can change the motor response to sensory feedback (Herzfeld et al, 2018;Ohmae and Medina, 2015;Romano et al, 2018;Ten Brinke et al, 2015;Yang and Lisberger, 2014). However, the role of cerebellar plasticity during tasks that include not only motor but also non-motor functions is still enigmatic (Chabrol et al, 2019;Deverett et al, 2018;Gao et al, 2018;Heffley et al, 2018;Hull, 2020;Kostadinov et al, 2019;Larry et al, 2019;Sendhilnathan et al, 2020;Tsutsumi et al, 2019). In particular, it is unclear when climbing fibers are activated during the acquisition of such tasks, to what extent they are causally linked to the related entrained movements as well as the expectation, presentation and/or omission of rewards, and what their impact is on concomitant simple spike modulation.…”

Section: Introductionmentioning

confidence: 99%

Complex spike firing adapts to saliency of inputs and engages readiness to act

Bina¹,

Romano²,

Hoogland

et al. 2020

Preprint

View full text Add to dashboard Cite

The cerebellum is involved in cognition next to motor coordination. During complex tasks, climbing fiber input to the cerebellum can deliver seemingly opposite signals, covering both motor and non-motor functions. To elucidate this ambiguity, we hypothesized that climbing fiber activity represents the saliency of inputs leading to action-readiness. We addressed this hypothesis by recording Purkinje cell activity in lateral cerebellum of awake mice learning go/no-go decisions based on entrained saliency of different sensory stimuli. As training progressed, the timing of climbing fiber signals switched in a coordinated fashion with that of Purkinje cell simple spikes towards the moment of occurrence of the salient stimulus that required action. Trial-by-trial analysis indicated that emerging climbing fiber activity is not linked to individual motor responses or rewards per se, but rather reflects the saliency of a particular sensory stimulus that engages a general readiness to act, bridging the non-motor with the motor functions.

show abstract

Cerebellar climbing fibers encode expected reward size

Cited by 29 publications

References 24 publications

Complex spikes encode reward expectation signals during visuomotor association learning

Complex spikes encode reward expectation signals during visuomotor association learning

Variability in error-based and reward-based human motor learning is associated with entorhinal volume

Complex spike firing adapts to saliency of inputs and engages readiness to act

Contact Info

Product

Resources

About