| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Neuroscience, Erasmus MC U, Rotterdam, The Netherlands
Submitted 9 November 2005; accepted in final form 10 December 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
). In this way, the flocculus is part of the eye reflex circuit involved in the control of ocular reflexes such as the vestibulo-ocular reflex (VOR) and the optokinetic reflex (OKR). These reflexes aid vision by maintaining a stationary projection of a visual scene on the retina on the basis of vestibular and visual motion signals, respectively.
Like in all other regions of the cerebellar cortex, all signals converge on Purkinje cells that deliver the whole output of the flocculus. Each Purkinje cell also receives prominent excitatory input from a single climbing fiber, originating from the contralateral inferior olive (IO). Activity of the climbing fiber (CF) elicits characteristic multipeaked complex spikes (CSs). This distinct input system has received much attention in the last decades for its potential teaching function in the formation of a cerebellar motor memory or its function in real-time motor control (Simpson et al. 1996).
The CSs of the flocculus are known to respond to visual stimulation, more specifically slip of the retinal image. This has been shown in the anesthetized rabbit using optokinetic stimulation of a large part of the visual field (Graf et al. 1988). Direct projections from motion sensitive visual nuclei such as the Accessory Optic System (AOS) and the Nucleus of the Optic Tract (NOT) to the dorsal cap (Giolli et al. 1985; Maekawa and Takeda 1977, 1979; Takeda and Maekawa 1976), which is the part of the IO that projects to the flocculus, can explain this. However, the dorsal cap does not only receive visual input. For instance the prepositus hypoglossi nucleus (PrH) is the main source of inhibitory projections to the dorsal cap of the IO (De Zeeuw et al. 1993, 1995; Frens et al. 2001). The PrH is thought to carry an efference copy of oculomotor commands (McCrea 1988; McFarland and Fuchs 1992). Unilateral and bilateral lesions of the PrH showed significant effects on the average floccular CS firing rate but had no apparent effect on CS modulation to constant velocity optokinetic stimulation (Arts et al. 2000). However, because these experiments were performed in anesthetized animals, it is very possible that no oculomotor signals could be relayed to the IO. Such projections make it unlikely that the CS input to the flocculus contains only sensory (i.e., retinal slip) information. It is the purpose of this study to specifically characterize an extraretinal motor component in the CF code.
Some indications that extraretinal signals related to self-motion influence CS modulation have been reported. Several groups showed that robust sinusoidal rotation of rabbits in complete darkness could produce residual CS modulation in a large fraction of floccular Purkinje cells (De Zeeuw et al. 1995; Ghelarducci et al. 1975; Simpson et al. 2002). In these experiments, it is impossible to relate this modulation exclusively to vestibular or oculomotor signals. Using transparent optokinetic stimulation, our laboratory recently showed that modulation of floccular CS under comparable retinal slip conditions differs with the oculomotor behavior of the rabbit (Frens et al. 2001). This strongly suggests an extraretinal influence, but the nature of such a modulation is presently unknown.
In monkey, (para-)floccular climbing fiber activity correlates better and fits more linearly with eye movement than with retinal slip during an ocular following task (Kobayashi et al. 1998). However, these authors chose to correlate the CS activity to the mean eye movement at 10 ms after the spike. Therefore their data may provide insight in a putative efferent motor consequence of the CS. Our study takes the opposite approach by asking what a Purkinje cell receiving a CS could infer about the past, present, or future oculomotor behavior of the animal.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Experiments were conducted using two female Dutch belted rabbits. Both animals were equipped for chronic experimentation with an acrylic head fixation pedestal and implanted search coils in both eyes. Recording chambers were positioned above the paramedian lobule of the cerebellum. All surgical procedures have been published elsewhere (Frens et al. 2000; Mathoera et al. 1999). Surgical procedures and experimental protocols are in accordance with the guidelines set by the Animal Welfare Committee of the Erasmus University as well as with the Principles of Laboratory Animal Care (National Institutes of Health).
Neuronal recording and spike sorting
CF activity was recorded in both Purkinje cell layer and molecular layer using standard extracellular recording techniques. Recordings were considered to be single units when a simple spike pause was present or, for molecular layer recordings, when the CS magnitude was prominent and constant and the interspike interval distribution characteristic for a single CF. The electrode signal was preamplified, band-pass filtered (100–3,000 Hz; CyberAmp 380, Axon Instruments), and sampled (25 kHz; Power 1401, Cambridge Electronic Design). The raw electrode signal was further processed off-line. Fifty Hertz hum and harmonics was removed off-line from the electrode signal. Possible spikes were identified by level detection and sorted using the first four principal components of the total spike wave set (Goossens et al. 2001). All spike waves in the set were aligned onto the first positive directed peak, and corrections were made to account for different numbers of complex and simple spikes when relevant. The spike times of identified CS were stored for further analysis.
Neuron characterization and selection
Well-isolated units in the flocculus were initially tested for preferred OKS direction using a handheld pattern followed by sinusoidal vertical axis (VA) OKS that served as an extra control. Climbing fibers projecting to zones 2 and 4 of the flocculus respond best to rotation of the visual field around the vertical axis (De Zeeuw et al. 1994). Units showing this preferred direction were selectively used for our analysis. The visual input to VA units was further analyzed using 1-ms light pulses delivered to each eye separately in a random manner (200- to 1,224-ms interflash time). The light-pulse stimulus induced an (often bimodal) CS transient that peaked at 38 ± 1 ms, followed by a short inhibition. This latency toward the CS peak was considered the minimal visual delay. From a total of 168 recorded units, 91 were classified as VA neurons. For a subset of 32 cells, the quality and duration of the recordings was sufficient for further analysis. Recording duration ranged from 253.9 to 3,243.9 s (mean = 848.2 s), with CS numbers ranging from 313 to 5,982 (mean = 1,075.7).
Visual stimulation and eye movement recording
After the initial cell characterization, trials of 10 min optokinetic stimulation were presented to the rabbit. The stimulus consisted of vertical axis rotation of a random dot pattern projected on a cone-shaped translucent screen that was placed over the animal. The velocity of rotation was driven by colored noise with a Gaussian distribution of stimulus velocities. Four different noise stimuli were used (Fig. 2A) with identical power spectra, except for a scaling factor. For all stimuli the mean velocity was 0°/s, and the variance was 1, 2, 4, and 8.5°/s. All these stimuli contain velocities that are well within the velocity tuning range of the accessory optic system (Soodak and Simpson 1988), which is the prime source for retinal slip signals in the vestibulo-cerebellum. After the optokinetic paradigm, background CF activity in the dark was recorded for as long as the isolation permitted. Complete darkness was secured by placing two black hemispheres over the eyes in addition to switching off the room lights.
|
). The noise level of the eye position recording was 
50" (SD). Eye and stimulus position signals were low-pass filtered (300 Hz; Axon CyberAmp 380) and sampled at 1 kHz. Eye and stimulus velocity were computed off-line by differentiation and application of a Gaussian smoothing filter (
= 15 ms). Data analysis
BEHAVIOR. All data analysis was performed using the Matlab software package (The Mathworks). Power spectra, transfer, and coherence function estimates of the OKS and oculomotor behavior were computed with Welch's averaged periodogram method using 75% overlapping 1,000 point sections and Hamming windows. Sections containing saccades rarely occurred and were automatically excluded using a manually adjustable velocity threshold. A window of 250 ms before and after a saccade was additionally excluded from analysis.
The optimal delay between the stimulus and the ensuing eye movement was detected by finding the time lag that resulted in the maximal cross-correlation between these two signals. Stimulus/eye velocity relationships were determined by first aligning the eye and stimulus velocity signal in time using the optimal delay, followed by calculation of mean eye velocities coinciding with stimulus velocities that were binned to 0.1°/s steps. The gain was computed as the ratio between these values. The gains to ipsilateral and contralateral movement were averaged to get the absolute gain.
SPIKE SIGNALS.
Spike-triggered averages were made by aligning retinal slip and eye velocity signals about the arrival times of all N complex spikes in a recording. Missing values caused by exclusion of saccades were ignored. Because of the randomness (Keating and Thach 1995) and ultra-low firing rate (1 Hz) of the CS signal, each CS was considered to encode an independent event. As a consequence, the CS signal was defined as a one-symbol code, leading us to calculate how informative a CS occurrence is for the Purkinje cell using the CS-conditional transmitted information. We used a method for estimating the transmitted information similar to that used by Optican and others (Optican et al. 1991; Optican and Richmond 1987; Richmond and Optican 1990), except that they used the stimulus-conditional information. In short: CS-conditional transmitted information T(cs;X) is defined as the Kullback-Leibler divergence (Kullback and Leibler 1951) between the CS-conditional stimulus probability density function P(X|cs) and the expected stimulus distribution P(X)
 |
where n is the number of all stimulus values on which the probability densities are estimated. If P(X) and P(X|cs) are identical, the transmitted information is 0. If a CS encodes only a single stimulus value (x), the gained information equals –log2 P(x). Probability density functions (pdf) were estimated using convolution of all n data points with a Gaussian kernel
 |
where u =
, 
x is the SD and h is the bandwidth. Because eye movement distributions tend to be bimodal rather than a unimodal normal (e.g., Fig. 3B), the SD of the stimulus distributions x was replaced with the sample interquartile range divided by 1.35. For normal distributions, this equals the SD. K(u) is the Gaussian kernel function
 |
|
(x) when the underlying distribution is normal (Silverman 1986)
 |
where m is the number of CS used to estimate the CS-conditional stimulus pdf.
The probability densities were determined for stimulus values that were 0.1°/s apart. Higher resolutions only slightly improved the information estimates. Because of the sparseness of data points in the tails of the distributions, probability density estimates of especially the CS-conditional stimulus distributions would be unreliable in the tails. Therefore we truncated the distributions at the 2.5 and 97.5% percentiles. We used a bootstrapping procedure to estimate the mean (Tb) and variance (Tb2) of the transmitted information when the stimulus and CS response are independent. For these estimates we used 105 artificial CS trains composed of randomly shuffled inter-CS intervals. The variance of the estimate of the transmitted information (T2) was bootstrapped from the CS-conditional stimulus values using 103 iterations. The bias-corrected information values were calculated as
 |
and information variances became
 |
The variance of the information peak times was estimated from the peak times of a bootstrapped set of 103 spike-triggered information estimates. Each of these estimates was based on a randomly resampled set (with replacement) of CS firing times.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
2 = 1, 2, 4, and 8.5°/s about an average velocity of 0°/s; Fig. 2A). These stimuli could induce eye movements through the OKR (Fig. 1D), and the frequency range extended high enough to effectively dissociate eye movement from instantaneous retinal slip (Fig. 1E). Behavior
Figure 2 shows the response of the animals to the stimuli (Fig. 2A). The gain of the oculomotor response, as estimated by the transfer function, was highest for the lower frequencies in all stimulus conditions (Fig. 2B). However, increased OKS power within a frequency bin had a negative effect on the oculomotor gain, which underlines the nonlinear nature of the OKR. The coherence function (Fig. 2C) shows what fraction of the oculomotor response is linearly related to the OKS. The coherence functions are similar for all stimuli, decreasing almost linearly with increasing frequency. Note that frequency components >12 Hz in the oculomotor response are unrelated to the OKS. The phase shift of the response was such that it could be best described by a group delay of 80 ms for all stimuli (Fig. 2, D and E). At this delay, the OKS that caused the eye movement, accounted for only 1% of the variation in the actual OKS, and consequently the ensuing slip. Therefore our stimulus accomplished an effective dissociation between instantaneous slip and eye movement. When adjusted for the delay, the eye movement response appeared as a nonlinear function of the OKS velocity (Fig. 2F). The average velocity gain of the animals (Fig. 2G) ranged from 0.3 for 3°/s OKS velocity or higher to 0.75 at the lowest OKS velocities.
If one compares the velocity gain functions of the four stimuli, it is important to note that the stimuli have different velocity distributions. However, in the range that is present in all stimuli (<4°/s), the gain functions are practically identical, although at low velocities (<1°/s), the stimuli with a high variance (4 and 8.5°/s) induce apparently lower gains. Most likely this is related to the nonlinear response of the optokinetic reflex to retinal slip and the integration time that it requires. Increasing the variance will decrease the average duration that a certain velocity is presented.
Spike-triggered average versus spike-triggered information
To explore the visuo-motor context in which CS occur, one can compute the cross-correlation between CS firing and slip and or eye velocity for different time lags using the classical spike-triggered average (STA). However, a correlation only describes the linear dependence between two variables, whereas the CS signal is tuned nonmonotonically to retinal slip with a peak response for the majority of CF at slip velocities 
1°/s (Barmack and Hess 1980; Kusunoki et al. 1990; Leonard et al. 1988; Simpson and Alley 1974), and the putative CS tuning to eye movement is unknown. It is important to notice that the tuning of the CS signal to a particular stimulus velocity x (i.e., slip or eye movement) can be inferred from the CS conditional stimulus probability P(X|cs) (Fig. 3, A and B, dashed curves) and the unconditional stimulus probability P(X) (Fig. 3, A and B, solid curves). Using Bayes rule, the ratio between the two probabilities multiplied by the average CS firing rate produces the tuning of the CS signal to the stimulus (Fig. 3, C and D). To be able to capture more than the linear dependence of the CS response to the stimulus input, we quantified how much all stimulus probabilities change after the occurrence of a CS (Fig. 3, A and B, from solid to dashed) by calculating the CS conditional transmitted information (see METHODS). This measures sufficiently (Kullback and Leibler 1951) how much the CS conditional stimulus probabilities differ from the unconditional (expected) stimulus probabilities. A sharper tuning of the CS to slip or eye movement results in a more informative CS (i.e., the CS demarcates a more specific feature of the stimulus). The unit of information is expressed in bits, where 1 bit indicates perfect discrimination between two states that were equally probable before any information was received. If a CS occurrence provides no information about slip or eye movement input, the transmitted information yields zero (bit). Because the transmitted information is independent of the range of stimulus values, we can use it to directly compare the CS tuning to slip and eye movement.
STAs as well as spike-triggered information graphs are summarized in Fig. 4. The slip signal averages of nearly all units showed a similar profile within each OKS condition (Fig. 4A). Qualitatively, for all stimuli, the average slip around a CS is directed toward the contralateral (nasal) direction before the spike and is followed by a peak directed toward the ipsilateral side at about the time of the spike. Because the delay between this second peak and the CS is shorter than the minimal visual delay [38 ± 1 (SE) ms; gray dashed line; see METHODS], only the contralateral directed slip can be responsible for the CS generation.
|
From the STA analysis it became clear that a CS is not only correlated with slip but also with eye movement. In Fig. 4, C and D, we compare the mean transmitted information curves for the different stimulus conditions. As one can see, the average curves of both the slip information and the eye velocity information are surprisingly independent of stimulus condition. This means that these profiles describe the behavior of the IO over a large range of OKS conditions.
When we compare the maximum information of the individual CF signals, we find that the information about the eye movement exceeds the slip information in all neurons (Fig. 5A). On average the eye movement information (in the order of stimuli with increasing variance: 0.61 ± 0.03, 0.75 ± 0.05, 0.65 ± 0.06, and 0.74 ± 0.06 bits, respectively) was more than twice the information of the slip (0.26 ± 0.02, 0.20 ± 0.03, 0.27 ± 0.01, and 0.21 ± 0.04 bits), irrespective of the stimulus used. There was no significant correlation between these two parameters (r = 0.14; P = 0.37). As one can see from Fig. 5A, it is impossible to qualify a CF as solely "sensory" or "motor," because there seems to be a continuum of possible sensori-motor combinations.
|
Spontaneous activity in the dark
To test the notion of a motor component in the CF code independently, we performed an additional experiment where eye movements were recorded in the absence of sensory stimulation.
In total darkness, eye position drifts spontaneously at low velocities. This type of drift is incorporated by the oculomotor integrator (Frens and Van Opstal 1994) and therefore accounted for by the oculomotor system. If there is indeed a direct contribution of oculomotor input to the CS generation, we expect that a significant information peak can still be observed if the eye movement is related to "spontaneous" CS activity while the animal sits in the dark. Obviously CS cannot provide information about retinal slip under these circumstances. Nor can an indirect correlation between the CS and the eye movement that is caused by a correlation of slip with both the CS and the eye movement play a role (see DISCUSSION). However, because of this latter fact and the suboptimal velocities of the drifting movements (µ 0°/s; 0.25°/s; cf. Fig. 5D), the amount of eye movement information is expected to be smaller than during oculomotor behavior in the light.
Because there is not much movement of the eyes in darkness, and because of the low firing rate of the CF, one requires substantial recording times to obtain sufficient data. Nonetheless, we could do this in nine different cells (mean recording time about 15 min). The results are shown in Fig. 6. The average drift of the eye in darkness was into the ipsilateral direction (Fig. 6, A and B, dashed line). In eight of nine cells, the STA shows a deviation from this drift velocity peaking toward the contralateral direction at about the time of the CS. Consequently, the occurrence of a CS increases the chance that the eye was moving contralaterally. The transmitted information that is associated with this deviation (Fig. 6C) resembles the eye velocity graphs that were obtained during OKS, with a highly similar timing of information (–7 ± 2 ms in darkness vs. –14 ± 1 on average during OKS), but—as expected—with a lower peak value (0.08 ± 0.01 bits). On the basis of these data, we conclude that part of "spontaneous" CS activity in the dark can be attributed to oculomotor activity.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
). The dependence on stimulus amplitude and frequency content are both incompatible with a linear system. A second disadvantage of sinusoidal stimulation is that it causes temporal correlation between instantaneous optokinetic stimulus velocity, retinal slip, eye velocity, and CS modulation. Without knowing the proper neural gains and delays of each signal, it is impossible to properly relate them to each other, without scanning a large range of frequencies. We set out to overcome this problem by presenting stimuli with lesser temporal structure, i.e., colored noise.
Using these stimuli, we showed that floccular CS firing is influenced by an oculomotor signal in addition to the retinal slip signal that was already recognized (Graf et al. 1988). As previously reported, the direction of the slip that elicited CS activity of VA Purkinje cells was toward the nasal side. The latency between the slip information peak and the CS was on average –131 ± 7 ms, which was slightly larger than the delay of 100 ms described in the monkey (Stone and Lisberger 1990). However the CS is not well time-locked to retinal slip (Fig. 5B). For the oculomotor signals, the optimal direction for evoking CS activity is in the same direction as the slip, but the timing is different, i.e., only a few milliseconds before the CS. In our paradigm, the height of the information peak about the ongoing eye movement was higher than the slip peak and more tightly coupled in time (Fig. 4, E and F). Also the CS tuning curves to eye velocity were considerably less variable (Fig. 5, C and D). Therefore it seems that the occurrence of a CS tells a Purkinje cell more about the oculomotor behavior of the animal than about the retinal slip. Note that this higher information can be partially caused by the cross-correlation with the visual input and partially because of direct motor efference copy, but that for a Purkinje cell the source of information is irrelevant.
Even in the absence of vision, an oculomotor signal was detectable in the CS activity, which independently shows a motor component in floccular CS (Fig. 6). Because the velocity of the spontaneous drifting movements is well below the optimal velocity to trigger a CS, the amount of information in darkness substantially lower (by a factor 10) than during OKS. However, the timing of the information is in complete agreement in both paradigms.
Neither during OKS nor in the dark did we observe systematic oculomotor activity after the CS. This shows that CS do not cause any oculomotor behavior. This is in contrast with previously reported data on the ocular following behavior (combined optokinetic and smooth pursuit responses) in the ventral paraflocculus of the monkey (Kobayashi et al. 1998). These authors used a stimulus that was highly correlated in time (i.e., constant speed), and could therefore not directly dissociate the slip and the eye velocity component. Their argument is based on a more linear relation between CS activity and eye velocity at 10 ms after the spike than between CS activity and slip velocity at 40 ms before the spike. However, if one compares these relations within the same velocity range (between 0 and 40°/s), the linearity seems equal (their Fig. 6, C and D). Only at higher slip velocities does the relation saturate. These higher velocities were not tested in the eye movement domain. Whether CSs result in eye movements in the monkey therefore remains an open question.
Although it is widely accepted that floccular complex spikes report the occurrence of retinal slip, we now show that, for a floccular VA Purkinje cell, a CS also provides information about a simultaneous eye movement. However, a correlation between eye movement and CS firing does not necessarily mean that this is caused by a direct causal relation. If an efference copy of the motor command is relayed to the IO, the motor command is the common ancestor that forms the causal link between CS and eye movement (Fig. 7, route R1). However, other routes that lead to a correlation between these two signals coexist. As mentioned above (Fig. 4, A and B), an instantaneous correlation between instantaneous slip and eye movement is merely caused by the definition of retinal slip (Fig. 7, route R2). For a temporal correlation between the CS and eye velocity, an alternative (although not mutually exclusive) explanation is that retinal slip is the common cause of both the CS and the compensatory eye movement (Fig. 7, route R3). Because of the correlation between slip and ensuing eye movement, a CS can provide information about both signals even if the eye movement is not causal to CS generation. However, it is highly unlikely that this latter option explains our results completely. Such a notion seems incompatible with the stronger relation that we find between eye velocity and CS, both in peak information (Fig. 5A) and timing (Fig. 5B). It would also predict a positive correlation between the amount of information that a CS encodes in the slip and the eye velocity domain. This is not the case (Fig. 5A). Finally, such a scheme could not explain the finding that also eye movements in the dark induce CF activity (Fig. 6). Our experiments cannot show the type (excitatory or inhibitory) of the anatomical projections that relay oculomotor information. It is impossible to distinguish excitation from lack of inhibition (or vice versa) in a combined signal.
|
; Ito 2001), and evidence is accumulating that other cerebellar plastic processes are also affected by climbing fiber activity (Hansel et al. 2001; Jorntell and Ekerot 2003). Retinal slip has been proposed as the teaching signal for floccular Purkinje cells, because it has been thought to represent visual error (Ito 1982). Through LTD, the synaptic input that can be associated with this error is gradually eliminated. However, situations exist where the flocculus should not interpret retinal slip as an error signal. Self-generated eye rotations over a stable visual background induce contralateral retinal slip, which in turn would induce CS firing. The oculomotor contribution that we find may serve to ensure that retinal slip is not relayed to the flocculus if extraretinal visual signals can predict the occurrence of that slip, because nothing needs to be learned. In this way the complex spike signal is refined to report visual perturbations that could not be anticipated and shows a remarkable analogy with blocking phenomena that have been shown in other experimental paradigms such as classical eye blink conditioning (Kim et al. 1998; Medina et al. 2002) or with gating of cutaneous sensory CF activity (Apps 1999). It should be noted that these results are in line with Marr-Albus-Ito models that require "motor error" for oculomotor learning. We propose a more sophisticated error signal, but the essentials of the learning mechanisms remain.
The concept that a sensory expectation inferred from a motor command is subtracted from actual sensory feedback dates back half a century (von Holst and Mittelstaedt 1950). The application to CF signals, qualifying the CSs as "unexpected event" or "error" messages, was thoroughly researched and debated during the last two decades (see for extensive review: Simpson et al. 1996). However, the retinal slip signal conveyed by floccular CF afferents was itself regarded as an error signal, and no distinction was made between expected and unexpected slip. The particular proposition that unexpected retinal slip could be actively gated by the IO has been hypothesized recently (Devor 2000, 2002). Here the term "expected" is still faintly associated with "voluntary," so that hypothetically only smooth pursuit and saccadic eye movements qualify to induce expected slip. We generalize the concept of expected slip to slip that could be inferred from head and eye movement signals.
Qualitatively, the oculomotor contribution as we measure it can indeed serve to create an "unexpected" slip signal. Eye movements induce slip in the direction opposite to the movement. The fact that the slip peak and the oculomotor peak have the same sign (both nasally directed) therefore fits the notion of "unexpected." A complicating factor may be that the slip and the oculomotor peak do not coincide temporally but are roughly 75 ms apart (Fig. 4F). However, it should be noted that we correlate the real slip and eye velocity to the CS (Fig. 7, left), whereas the IO receives neural correlates of these signals (Fig. 7, right). Consequently, the actual signals that arrive in the olive may be more synchronous.
In this study, we have limited the analysis to retinal slip velocity and eye movement velocity. This does not mean that other modalities, or other orders of the same modalities, could be present in the CF code. For instance Simpson et al. (2002) reported a vestibular component, and Kobayashi et al. (1998) reported a small but significant correlation with acceleration.
The oculomotor signal makes CS firing more likely when the eyes rotate toward the contralateral direction. This is different from results found previously (Simpson et al. 2002), where a residual CS modulation to vestibular stimulation in the absence of vision (VOR dark) was shown. The oculomotor signal described in our study is unlikely to be the same signal that causes the modulation under VOR dark conditions, because the CS facilitating eye movement directions are opposite. During VOR in the dark, complex spikes tend to occur when the eye moves toward the ipsilateral side and the head toward the contralateral side. This suggests that vestibular signals also play a role in CS modulation and counteract the influence of the oculomotor component. Strikingly, this is exactly what one would expect, because expected slip caused by self-motion can be estimated by the difference between head velocity (vestibular) and eye velocity (oculomotor).
Because of the 100% efficiency of the CF-Purkinje cell synapse, the source of extraretinal signals present in floccular CS activity lies at the level of the IO or further upstream and remains to be established. Given the timing of the oculomotor signal (at about 14 ms before the CS), we expect the source of the signal to be an efference copy of the oculomotor command rather than a sensory registration of the actual movement. The latency is simply too short to allow for a sensory involvement. The aforementioned projection from the NPH to the IO seems a likely candidate, but further research is necessary to establish such a role.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: M. A. Frens, Dept. of Neuroscience, Erasmus MC, PO Box 1738, 3000 DR Rotterdam, The Netherlands (E-mail: m.frens{at}erasmusmc.nl)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Arts MP, De Zeeuw CI, Lips J, Rosbak E, and Simpson JI. Effects of nucleus prepositus hypoglossi lesions on visual climbing fiber activity in the rabbit flocculus. J Neurophysiol 84: 2552–2563, 2000.
Barmack NH and Hess DT. Multiple-unit activity evoked in dorsal cap of inferior olive of the rabbit by visual stimulation. J Neurophysiol 43: 151–164, 1980.
Bear MF and Linden DJ. The mechansims and meaning of long-term synaptic depression in the mammalian brain. In: Synapses, edited by Cowan WM, Südhof TC, and Stevens CF. Baltimore, MD: Johns Hopkins University Press, 2000, p. 455–517.
Blazquez PM, Hirata Y, and Highstein SM. The vestibulo-ocular reflex as a model system for motor learning: what is the role of the cerebellum? Cerebellum 3: 188–192, 2004.[CrossRef][Web of Science][Medline]
Collewijn H. Optokinetic eye movements in the rabbit: input-output relations. Vision Res 9: 117–132, 1969.[CrossRef][Web of Science][Medline]
Devor A. Is the cerebellum like cerebellar-like structures? Brain Res Brain Res Rev 34: 149–156, 2000.[CrossRef][Medline]
Devor A. The great gate: control of sensory information flow to the cerebellum. Cerebellum 1: 27–34, 2002.[CrossRef][Medline]
De Zeeuw CI, Wentzel P, and Mugnaini E. Fine structure of the dorsal cap of the inferior olive and its GABAergic and non-GABAergic input from the nucleus prepositus hypoglossi in rat and rabbit. J Comp Neurol 327: 63–82, 1993.[CrossRef][Web of Science][Medline]
De Zeeuw CI, Wylie DR, DiGiorgi PL, and Simpson JI. Projections of individual Purkinje cells of identified zones in the flocculus to the vestibular and cerebellar nuclei in the rabbit. J Comp Neurol 349: 428–447, 1994.[CrossRef][Web of Science][Medline]
De Zeeuw CI, Wylie DR, Stahl JS, and Simpson JI. Phase relations of Purkinje cells in the rabbit flocculus during compensatory eye movements. J Neurophysiol 74: 2051–2064, 1995.
Frens MA, Mathoera AL, and Van der Steen J. On the nature of gain changes of the optokinetic reflex. Prog Brain Res 124: 247–255, 2000.[Medline]
Frens MA, Mathoera AL, and van der Steen J. Floccular complex spike response to transparent retinal slip. Neuron 30: 795–801, 2001.[CrossRef][Web of Science][Medline]
Frens MA and Van Opstal AJ. Auditory-evoked saccades in two dimensions: dynamical characteristics, influence of eye position and sound source spectrum. In: Information Processing Underlying Gaze Control, edited by Delgado-Garcia JM, Vidal P, and Godaux E. New York: Pergamon, 1994, p. 328–341.
Ghelarducci B, Ito M, and Yagi N. Impulse discharges from flocculus Purkinje cells of alert rabbits during visual stimulation combined with horizontal head rotation. Brain Res 87: 66–72, 1975.[CrossRef][Web of Science][Medline]
Giolli RA, Blanks RH, Torigoe Y, and Williams DD. Projections of medial terminal accessory optic nucleus, ventral tegmental nuclei, and substantia nigra of rabbit and rat as studied by retrograde axonal transport of horseradish peroxidase. J Comp Neurol 232: 99–116, 1985.[CrossRef][Web of Science][Medline]
Goossens J, Daniel H, Rancillac A, van der Steen J, Oberdick J, Crepel F, De Zeeuw CI, and Frens MA. Expression of protein kinase C inhibitor blocks cerebellar long-term depression without affecting Purkinje cell excitability in alert mice. J Neurosci 21: 5813–5823, 2001.
Graf W, Simpson JI, and Leonard CS. Spatial organization of visual messages of the rabbit's cerebellar flocculus. II. Complex and simple spike responses of Purkinje cells. J Neurophysiol 60: 2091–2121, 1988.
Hansel C, Linden DJ, and D'Angelo E. Beyond parallel fiber LTD: the diversity of synaptic and non-synaptic plasticity in the cerebellum. Nat Neurosci 4: 467–475, 2001.[Web of Science][Medline]
Ito M. Cerebellar control of the vestibulo-ocular reflex–around the flocculus hypothesis. Annu Rev Neurosci 5: 275–296, 1982.[CrossRef][Web of Science][Medline]
Ito M. Cerebellar long-term depression: characterization, signal transduction, and functional roles. Physiol Rev 81: 1143–1195, 2001.
Jorntell H and Ekerot CF. Receptive field plasticity profoundly alters the cutaneous parallel fiber synaptic input to cerebellar interneurons in vivo. J Neurosci 23: 9620–9631, 2003.
Keating JG and Thach WT. Nonclock behavior of inferior olive neurons: interspike interval of Purkinje cell complex spike discharge in the awake behaving monkey is random. J Neurophysiol 73: 1329–1340, 1995.
Kim JJ, Krupa DJ, and Thompson RF. Inhibitory cerebello-olivary projections and blocking effect in classical conditioning. Science 279: 570–573, 1998.
Kobayashi Y, Kawano K, Takemura A, Inoue Y, Kitama T, Gomi H, and Kawato M. Temporal firing patterns of Purkinje cells in the cerebellar ventral paraflocculus during ocular following responses in monkeys II. Complex spikes. J Neurophysiol 80: 832–848, 1998.
Kullback S and Leibler RA. On information and sufficiency. Ann Math Stat 22: 79–86, 1951.[CrossRef]
Kusunoki M, Kano M, Kano MS, and Maekawa K. Nature of optokinetic response and zonal organization of climbing fiber afferents in the vestibulocerebellum of the pigmented rabbit. I. The flocculus. Exp Brain Res 80: 225–237, 1990.[CrossRef][Web of Science][Medline]
Leonard CS, Simpson JI, and Graf W. Spatial organization of visual messages of the rabbit's cerebellar flocculus. I. Typology of inferior olive neurons of the dorsal cap of Kooy. J Neurophysiol 60: 2073–2090, 1988.
Maekawa K and Takeda T. Afferent pathways from the visual system to the cerebellar flocculus of the rabbit. In: Control of Gaze by Brain Stem Neurons, edited by Baker R and Berthoz A. Amsterdam: Elsevier/North-Holland, 1977, p. 187–195.
Maekawa K and Takeda T. Origin of descending afferents to the rostral part of dorsal cap of inferior olive which transfers contralateral optic activities to the flocculus. A horseradish peroxidase study. Brain Res 172: 393–405, 1979.[CrossRef][Web of Science][Medline]
Mathoera AL, Frens MA, and van der Steen J. Visual-vestibular interaction during transparent optokinetic stimulation in the rabbit. Exp Brain Res 125: 87–96, 1999.[CrossRef][Web of Science][Medline]
McCrea RA. Neuroanatomy of the oculomotor system. The nucleus prepositus. Rev Oculomot Res 2: 203–223, 1988.[Medline]
McFarland JL and Fuchs AF. Discharge patterns in nucleus prepositus hypoglossi and adjacent medial vestibular nucleus during horizontal eye movement in behaving macaques. J Neurophysiol 68: 319–332, 1992.
Medina JF, Nores WL, and Mauk MD. Inhibition of climbing fibres is a signal for the extinction of conditioned eyelid responses. Nature 416: 330–333, 2002.[CrossRef][Medline]
Optican LM, Gawne TJ, Richmond BJ, and Joseph PJ. Unbiased measures of transmitted information and channel capacity from multivariate neuronal data. Biol Cybern 65: 305–310, 1991.[CrossRef][Web of Science][Medline]
Optican LM and Richmond BJ. Temporal encoding of two-dimensional patterns by single units in primate inferior temporal cortex. III. Information theoretic analysis. J Neurophysiol 57: 162–178, 1987.
Richmond BJ and Optican LM. Temporal encoding of two-dimensional patterns by single units in primate primary visual cortex. II. Information transmission. J Neurophysiol 64: 370–380, 1990.
Silverman BW. Density Estimation for Statistics and Data Analysis. London: Chapman and Hall, 1986.
Simpson JI and Alley KE. Visual climbing fiber input to rabbit vestibulo-cerebellum: a source of direction-specific information. Brain Res 82: 302–308, 1974.[CrossRef][Web of Science][Medline]
Simpson JI, Belton T, Suh M, and Winkelman B. Complex spike activity in the flocculus signals more than the eye can see. Ann NY Acad Sci 978: 232–236, 2002.[CrossRef][Web of Science][Medline]
Simpson JI, Wylie DR, and De Zeeuw CI. On climbing fiber signals and their consequence(s). Behav Brain Sci 19: 380–394, 1996.
Soodak RE and Simpson JI. The accessory optic system of rabbit. I. Basic visual response properties. J Neurophysiol 60: 2037–2054, 1988.
Stone LS and Lisberger SG. Visual responses of Purkinje cells in the cerebellar flocculus during smooth-pursuit eye movements in monkeys. II. Complex spikes. J Neurophysiol 63: 1262–1275, 1990.
Takeda T and Maekawa K. The origin of the pretecto-olivary tract. A study using the horseradish peroxidase method. Brain Res 117: 319–325, 1976.[CrossRef][Web of Science][Medline]
Van der Steen J and Collewijn H. Ocular stability in the horizontal, frontal and sagittal planes in the rabbit. Exp Brain Res 56: 263–274, 1984.[Web of Science][Medline]
Von Holst E and Mittelstaedt H. Das reafferenzprinzip (The principle of reafference). Naturwissenschaften 37: 464–476, 1950.[CrossRef][Web of Science]
This article has been cited by other articles:
|
|
 |
|
  T. Liu, D. Xu, J. Ashe, and K. Bushara Specificity of Inferior Olive Response to Stimulus Timing J Neurophysiol, September 1, 2008; 100(3): 1557 - 1561. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Kawato From 'Understanding the Brain by Creating the Brain' towards manipulative neuroscience Phil Trans R Soc B, June 27, 2008; 363(1500): 2201 - 2214. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. H. Barmack and V. Yakhnitsa Functions of Interneurons in Mouse Cerebellum J. Neurosci., January 30, 2008; 28(5): 1140 - 1152. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. S. Kim, S. Y. Moon, K. -D. Choi, J. -H. Kim, and J. A. Sharpe Patterns of ocular oscillation in oculopalatal tremor: Imaging correlations Neurology, April 3, 2007; 68(14): 1128 - 1135. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
