Electrophysiological study of cerebellar
fastigial nuclei in the eye blink
conditioning.
Milen Chernev Angelov
Research Master’s - Brain and Cognitive Sciences
University of Amsterdam
Summary:
Delay eyeblink conditioning (DEC) is an ideal paradigm for studying
circuit mechanisms of associative learning. The cerebellar simplex lobe
and its downstream anterior interposed nucleus are the key regions
mediating conditioned, but not unconditioned responses. However, the
involvement of other cerebellar circuits remains unclear. Here we
demonstrate the electrophysiological patterns of fastigial nucleus (FN) in
controlling eyeblink responses. FN glutamatergic neurons predict the
amplitude of eyelid closures. Our data highlight the role and importance
of the FN in modulating conditioned and unconditioned behavior in
mice.
Introduction:
Sensorimotor associative behavior allows vertebrates to convert perceptions of the environment into specific motor executions. Pavlovian DEC is an ideal model for studying the cellular and circuit mechanisms for the control of associative tasks in which the motor response is precisely timed with respect to the sensory input (Gormezano et al., 1962; McCormick and Thompson, 1984; Patterson et al., 1977). In this paradigm, animals are presented with a neutral conditioned stimuli (CS)
followed by an unconditioned stimuli (US) that reliably causes an unconditioned eyeblink reflex (UR). In the case of the current study, the CS is a green light placed in front of the animal and the US is an air puff to the eye. Prior to conditioning, the CS does not elicit any motor output. After conditioning, the CS is associated with the US and produces a well-timed conditioned eyeblink response (CR) that reaches its maximum amplitude at the moment when the US is about to occur.
Previous studies unambiguously reveal the crucial involvements of the cerebellum and its circuits in controlling DEC (McCormick & Thompson, 1984; Medina et al., 2000; Gao et al., 2016). This is evident for both acquisition and
expression of DEC (De Zeeuw & Yeo, 2005; McCormick and Thompson, 1984). The well-defined modular topographical circuitry of the cerebellum provides a unique entry for studying the contributions of its specific cortical and nuclear regions to sensorimotor tasks. Studies over the past decades have provided paramount evidence for the roles of the simplex lobule in the cerebellar hemispheres and the downstream anterior interposed nuclei (IN) in DEC (Chen et al., 1999; Heiney et al., 2014b; Koekkoek et al., 2003; Mauk et al., 2014; Mostofi et al., 2010; Steinmetz and Freeman, 2014; ten Brinke et al., 2015; Ten Brinke et al., 2017). These regions receive both mossy fiber (MF) and climbing fiber (CF) inputs (Heckroth and Eisenman, 1988; Van der Want et al., 1989), which relay the CS and US signals, respectively (Jirenhed et al., 2017; Johansson et al., 2016; Ohmae and Medina, 2015; ten Brinke et al., 2015). Based upon their activity, various forms of synaptic and structural plasticity could occur during DEC (Boele et al., 2013; Carey, 2011; Koekkoek et al., 2003; Ten Brinke et al., 2017), resulting in suppression of simple spike activity in Purkinje Cells (PC), which in turn disinhibits the neurons in IN, and eventually drives the conditioned, but not unconditioned, eyelid closure via the premotor red nucleus (RN) and facial motoneurons (Koekkoek et al., 2003; Mauk et al., 2014).
At present it is unclear to what extent other cerebellar modules contribute to DEC. The vermis projects to the fastigial nucleus (FN), which targets vast numbers of
downstream brain regions (Batton et al., 1977). Indeed, FN outputs have recently been implied to play various roles in both motor and non-motor tasks (Gao et al., 2018; Yamada et al., 2019; Zhang et al., 2016). Anatomical studies using retrograde trans-neuronal tracing of rabies virus from the eyelid muscle orbicularis oculi have not only revealed prominent labeling in IN but also in FN, suggesting a potential role of vermis and FN in controlling eyelid movements (Gonzalez-Joekes and Schreurs, 2012; Morcuende et al., 2002). Yet, physiological evidence for the involvement of the FN in DEC is lacking.
Here we uncovered that FN neurons show CS related modulation, which strongly correlates with eyelid closure amplitudes on a trial-by-trial basis. DEC related modulations were observed exclusively in excitatory FN neurons. Interestingly, with optogenetic tagging two cell types (GABAergic and
glutaminergic), we found only excitatory cells are involved in controlling CR and UR.
Method:
Surgical procedures
Mice were anesthetized with 5% isoflurane for induction and 2.5% for
maintenance. Animals were fixed on a mouse stereotaxic surgical plate (David Kopf Instruments). Eyes were covered with DuraTears (Alcon Laboratories, Inc.), and body temperature was kept at 37 ± 0.5 oC constantly during the operation. We
injected bupivacaine (4mg/kg) intraperitoneally after surgery.
For electrophysiology, after removing hair over the skull, we sprayed lidocaine (2.5 mg/mL) locally on the skin, and a vertical skin cut was applied to expose skull. The skull was pre-treated with Optibond All-in-one (Kerr), and a 5.5 x 4.0 mm custom made pedestal was attached to the skull with Charisma (Heraeus Kulzer). Mice were allowed to recover at least two weeks after surgery and prior to behavior studies. Before electrophysiology experiments, a small craniotomy (diameter = 2.0 mm) was performed on the skull over recording targets. We built a chamber around the craniotomy with Charisma and sealed with Picodent twinsil.
For intracranial viral injections, anesthetized animals were fixed on a mouse
stereotaxic surgical plate. The skull was exposed and the head was positioned so that the bregma and lambda were levelled. The coordinates for the FN location are AP – 2.7 (mm), ML – 0.8 (mm) Depth – 2.4 (mm) (The origin of coordinates is defined as the anterior tip of the interparietal bone. Viral injection volume was 50 nl). We gently lowered a capillary (⌀ = 8 μm opening) and AAV viral vectors were slowly injected
in the targeted regions. Glass capillaries were left at the injection sites for > 5 mins before slowly retracted from the brain. To optogenetically tag FN neurons in awake animals, a 2 mm long optical fiber (⌀ = 200 μm, 0.22 NA, ThorLabs) was inserted
through a small craniotomy (⌀ = 300 μm) and chronically fixed on the skull with
Charisma.
Behavioral training
All mice were in-house bred in the Experimental Animal Center of Erasmus MC without exposure to any behavioral test, and considered ‘naive’ prior to our studies. Mice were head-fixed and suspended on a cylindrical treadmill in a light and sound-attenuated chamber. A light-emitting diode was present about 7 cm in front of the animals as the CS, together with a corneal air puff (tip opening ⌀ = 1.5
mm, 30 psi) placed about 1 cm to the left eye as the US. In a paired trial, the CS was 250 ms long and co-terminated with a 10 ms US. Eyelid movement was real-time monitored by a 250 fps camera (scA640-120gc, Basler) connected with NI-PXI system (National Instruments) which was also used to control the initiation and termination of triggers. Behavior data was acquired and digitized through a RHD2000
Evaluation System (Intan Technology) at 20 kHz sampling rate. Mice were trained with 200 paired trials with randomized inter-trial interval of 10-15 s for 7-10 days.
In-vivo electrophysiology
We used a multi-channel electrophysiological acquisition system in the studies. Electrodes were manipulated by a SM7 keypad controller (Luigs & Neumann) in XYZ axes. We recorded FN neurons at a depth of 2.0-2.7 mm, as
measured from the cerebellar surface. Neuronal signals were notch filtered at 50 Hz, amplified and digitized at 20 kHz sampling rate by Axon acquisition system
(1440A，Molecular Devices Corporation). Multi-channel recordings (32-channles ASSY-32-E2 or 64-channles ASSY 77H-H2, Cambridge NeuroTech) were amplified and digitized on an Intan RHD2000 Evaluation System (Intan Technology) at 20 kHz sampling rate, and were further analyzed offline using custom written codes in Matlab. In all electrophysiological experiments combined with DEC, at least 50 CS-US paired trials were given to animals, and neurons with a minimum of 20 CR trials were included in the datasets. For optogenetic studies, at least 10 trials with
optogenetic stimulation were delivered to investigate the cell responses to corresponding optogenetic perturbations.
Optogenetic Manipulation
Electrophysiological recordings were carried out four weeks after AAV injection and/or optical fiber implantation. We used orange LED light source (M595F2, Thorlabs) to activate ChrimsonR and blue LED light source (M470F3, Thorlabs) to activate ChR2. Optical lights with 100 Hz pulse, 50/50 duty cycle were control by a high power light driver (DC2100, Thorlabs). The whole optical fiber was wrapped with light-isolating aluminium foil, so that mice would not perceive the optogenetic light as a CS. To express ChrimsonR in the excitatory or inhibitory subgroups of FN neurons, AAV9-Syn-FLEX-ChrimsonR-tdTomato was injected in the FN of VGluT2-Cre or Gad2-Cre mice. To identify the ChrimsonR-expressing neurons, we first illuminated the orange light (125 ms, 4.5 mW) and recorded the neuronal responses in FN. Only neurons showing short latency responses to the optogenetic stimulation (latency < 20 ms) were included in the datasets for further analysis of their responses during behavior (Klapoetke et al., 2014; Mardinly et al., 2018).
Histology and microscopy
Animals were deeply anesthetized with intraperitoneal injection of
pentobarbital sodium solution (50 mg/kg, i.p.) and perfused transcardially with saline, followed by 4% paraformaldehyde (PFA)-0.1 M phosphate buffer (PB, pH 7.4). Brains were removed immediately and post-fixed overnight in 4% PFA-0.1 M PB at 4 oC. Fixed brains were placed in 10% sucrose overnight at 4 oC and embedded
in 12% gelatin-10% sucrose. After fixation in 10% formalin-30% sucrose overnight at 4 oC, serial coronal sections were cut with microtone (SM2000R, Leica) at 40 µm and
collected in 0.1 M PB. For immunohistochemistry and immunofluorescence purposes, sections were incubated subsequently with primary and secondary antibodies. For VGluT2 staining, guinea pig anti-VGluT2 primary antibody (1:750, Sigma-Aldrich, AB2251-I) and Alexa fluor® 488 donkey anti-guinea pig secondary antibody (1:400, Jackson, 706-545-148) were used. For Gad2 staining, rabbit anti-Gad65/67 primary antibody (1:1000, Sigma-Aldrich, AB1511), and Alexa fluor® 488 donkey anti-rabbit secondary antibody (1:400, Jackson, 711-545-152) were used. All antibodies were titrated for working solution with 2% normal horse serum-0.4% triton-0.1M PBS solution. Primary antibodies were incubated at 4 oC overnight, and
secondary antibodies were incubated at room temperature for 2 h. After each
incubation session, sections were gently rinsed with 0.1 M PBS (10 min, 3 times). For fluorescence imaging, we took overviews of the brains with a 10x objective on a
fluorescence scanner (Axio Imager 2, ZEISS) or high-magnification images on a confocal microscope (LSM 700, ZEISS). Images were post-processed with ImageJ and Adobe Illustrator.
Behavioral analysis
Eyelid data was firstly down-sampled to 1 kHz. To investigate eyelid position changes over CS and US, each trial was normalized to a 500 ms baseline prior to the CS onset. We removed trials with noisy baseline (spontaneous blinking) by
performing an iterative Grubbs’ outlier detection test (α = 0.05) on the standard deviations of baseline. A CR-trial was determined if eyelid closure exceeded 5% of mean baseline, and the CR onset was defined as the timing which eyelid closure exceeded 3 SDs of the baseline value. Peak amplitudes of CR and UR were detected in a 50-250 ms window during CS-US interval and a 100 ms window after US, respectively.
Electrophysiological analysis
Neurons from multi-channel recordings were sorted with JRCLUST (Jun et al., 2017), and all spike time was stored for further analysis. In order to determine the cell modulations during CR, only cells with more than 20 CR trials were included in the dataset. Peristimulus time histograms (PSTHs) of well-isolated units were constructed by superimposing CS-onset aligned spike time in a 50 ms, 5 ms-step shifting window, and were expressed as frequency. We conservatively shorten CR modulation detective window to 50-200 ms due to the fact that shifting window went across CR and UR in the last 50 ms of CS. Baseline firing rates were calculated as the mean frequency over the 500 ms window prior to the onset of CS. We
determined firing rate change during CS by subtracting the baseline firing rates with frequency within 50-200 ms. Cells with CS related firing rate changes over 3 SDs of the baseline frequency in 50-200 ms were considered as CS modulating neurons. We then discriminated the overall direction of modulation by a linear fit of spike
frequency in the 0-200 ms window. Fittings with positive slopes were considered as facilitation cells, whereas with negative slopes were considered as suppression cells. UR modulations were determined within the first 100 ms after US. Spike rate change over 5 Hz was chosen as the threshold for UR modulation, as previously described (Ten Brinke et al., 2017).
Statistics
All statistics were performed by using MATLAB and GraphPad Prism. Behavior was illustrated as average of all trials ± standard deviation. The neuron frequency changes were plotted as the average of all cells ± coefficient variant, and sample sizes were displayed in the figure legends. Statistical comparisons were performed by using t-tests, repeated-measures ANOVA and restricted maximum likelihood model depending on the experiment and data specificity, unless stated otherwise. Statistical significance was defined as p < 0.05, and annotations were *p < 0.05, **p < 0.01, ***p < 0.001 respectively. No significant difference was denoted as n.s.
Results:
Task-related Modulation of FN Neurons during DEC
Head-restrained mice were presented with a 250 ms green light cue as the CS, co-terminating with a 10 ms aversive periocular air puff as US (Figure 1A).
Following 7-10 consecutive days of training, expert mice responded to the CS with a well-timed CR, prior to the onset of UR (Figure 1B). To examine the involvement of FN in DEC, we measured the activity of FN neurons ipsilateral to the trained eye by recording from well-isolated single units in the expert mice (Figure 1C). FN neurons (n = 162) had diverse modulation patterns in response to CS (Figure 1D) and US (Figure 2. A to C). The majority of FN neurons increased their firing rates in response to CS (termed as facilitation cells, n = 86/162), by 64.4 ± 38.9 % of the baseline firing rates. A minor portion of FN neurons decreased their firing rates in response to CS (suppression cells, n = 16/162), with an average suppression of 33.2 ± 21.2 %. Both CS related facilitation and suppression neurons had clear US related modulation (Figure 2. A & B). In particular, discrete modulation features were found in the neurons with both CS and US related facilitation (p < 0.001; Figure 2. D). About 37% of FN neurons did not show any significant modulation during CS-US interval (no modulation cells, n = 60/162), while they presented a weaker modulation to the UR (Figure 1. E).
We next asked what kind of information these cells encode. It is possible that CS related modulation of FN neurons reflects sensory responses to the light, or alternatively, that the FN activity is encoding the CR performance of the DEC task. We therefore sought to further clarify the specific relationship between FN neuron activity and the behavioral outcome. For all CS responsive FN neurons, we analysed the trial-by-trial relationship between the magnitudes of neuron modulation and the
amplitudes of CR (Ten Brinke et al., 2017). Positive trial-by-trial correlations between the FN neuron modulation and the CR peak amplitudes were detected in a group of facilitation neurons ((P < 0.05, linear regression, n = 9 units; Figure 1. E). In other words, the modulation amplitudes of FN neurons could faithfully encode the amplitudes of the CR response at the single trial level. Interestingly, we found a similar portion of FN neuronsin which their CS related modulation negatively correlated with the CR response (P < 0.05, linear regression, n = 5 units; Figure 1. G &
H) suggesting complex encoding mechanisms in FN neurons.
To determine the temporal relationship between the FN activity and CR performance we used a three-dimensional correlation matrix. We analyzed the trial-by-trial correlation between FN neuron activities with the eyelid position at various time points throughout the task. The significant correlation between FN facilitation and CR performance were found above the diagonal line within the CS-US interval, revealing that the across-trial correlations were strongest when FN modulation preceded the eyelid closure (Figure 1. I). The peak correlation was found when FN facilitation occurred 40 ms prior to the CR (Figure 1. I). In line with this finding, the timing for the onset as well as peak of the FN facilitation was significantly earlier than that for the CR onset (P < 0.001, paired t-test; Figure 1. J) and CR peak (P < 0.01, paired t-test; Figure 1. K), respectively. In contrast, FN neurons with CS related suppression had minimal trial-by-trial correlation with CR performance (Figure 3. A to C). Yet, the timing for the onset and that for the trough of suppression were also significantly earlier than those for CR onset (P < 0.01, paired t-test; Figure 3. D) and peak (P < 0.001, paired t-test; Figure 3. E), respectively. Across all FN neurons with CS-related modulation, less than 3% of neurons had modulation onsets within 50 ms following CS (Figure 1. J). Therefore, it is unlikely that FN modulation represents a simple sensory response to the CS light. Taken together, these results reveal a clear correlation between FN activity and the motor performance of DEC, especially in the facilitation cells.
Glutamatergic neurons heavily recruited during DEC
Cerebellar nuclei comprise heterogeneous groups of neurons. In general, the large excitatory neurons are glutamatergic and project to diverse extra-cerebellar regions, whereas the GABAergic inhibitory neurons project mainly to the inferior olive and/or to local FN (inter)neurons (Uusisaari and Knopfel, 2012). To clarify which type(s) of FN neurons are engaged in DEC, we expressed ChrimsonR in either the excitatory or inhibitory neurons by stereotaxically injecting AAV9-Syn-FLEX-ChrimsonR-tdTomato into the FN of VGluT2-ires-Cre or Gad2-ires-Cre mice (Figure 4. A). ChrimsonR expressing FN neurons showed robust short latency facilitation
(7.1 ± 4.2 ms for VGluT2 and 5.0 ± 4.2 ms for Gad2) in response to optic activation (Figure 4. B & C, Figure 5. A & B), and therefore could be ‘opto-tagged’ as
glutamatergic or GABAergic FN neurons. In totally, 15 glutamatergic cells and 8 GABAergic cells were included in the dataset (Figure 5. C & D). When we recorded the CS related modulation patterns of the ‘opto-tagged’ excitatory FN neurons in well-trained mice, we found that they showed both CS related facilitation (average facilitation peak 195.2 ± 56.2%, n = 7/15) and suppression (average suppression trough 67 ± 28.8%, n = 3/15, Figure 4. D). In contrast, no modulation was found in any GABAergic neuron, which was significantly lower than the chance level of detecting a modulating neuron in FN (p = 3.08x10-4; Figure 4. E). Therefore, it is likely that the glutamatergic neurons were selectively recruited in DEC.
Discussion
:
In this study we provide evidence how FN neurons of the cerebellum present well-timed spike modulation in response to both CS and US. It is in particular the excitatory neurons of the FN that modulate well during DEC, allowing prediction of the CR response on a trial-by-trial basis. These data highlight that conditioned and unconditioned, sensorimotor behavior can be controlled by the cerebellum in a distributed, yet synergistic manner.
FN Activation is Essential for Blinking during DEC
Most of the task related FN neurons in the present study increased their activity during DEC. This finding is in line with the fact that CR amplitudes tend to correlate with the simplex lobule-IN module (ten Brinke et al., 2015). It is compatible with the concept that long-term potentiation (LTP) at the parallel fiber to molecular layer interneuron synapse and long-term depression (LTD) at the parallel fiber to PC synapse can synergistically control cerebellar motor teaching (Gao et al., 2012). In this respect, it would be curious to examine whether the proportion of FNs showing facilitation during DEC mismatches that of the vermal PCs with simple spike
suppression during DEC. If that is the case, several factors may contribute to this finding: First, this may point towards the possibility that it is actually only the PCs in only specific vermis lobules that determine the DEC related activity in FN. Second, sensorimotor information conveyed by the excitatory, mossy fiber and climbing fiber collaterals may also directly facilitate activity in FN during DEC (Medina et al., 2000; Najac and Raman, 2017; Ohmae and Medina, 2015). Hence, it is possible that inputs
from specific vermal PCs and/or pre-cerebellar mossy fiber and climbing fiber
sources contribute to the relatively dominant facilitation of FN neurons during DEC.
Multi-modular Control of Sensorimotor Tasks and Functional Implications
The current data highlight that the activity of different cerebellar modules can be modulated simultaneously during an associative learning task. How the different cerebellar modules are coordinated during such a sensorimotor task is unclear. The shared neural dynamics between the FN and IN module suggest that common inputs might to a certain extent facilitate the synergy across different functional modules. Mossy fibers carrying CS inputs ascend into cerebellar granule cells and could expansively split their parallel fibers across multiple cortical regions (Biswas et al., 2019; Na et al., 2019). Thus, similar task-related PC modulations in the simplex lobule and vermis could result from uniform mossy fiber inputs diverging in distinct cerebellar modules. In addition, the cerebellar nucleo-cortical feedback loops to the granule cell layer might also play a role in distributing the CS-related inputs (Gao et al., 2016). Indeed, also during DEC granule cells show dense representations of predictive signals, which could be driven in part by the excitatory nucleo-cortical projections from the CN neurons (Giovannucci et al., 2017).
In contrast to the MF projections that are widely distributed across different cerebellar modules, the CFs derived from distinct inferior olivary subnuclei provide highly discrete inputs to specific modules (Ruigrok, 2011). During DEC, the climbing fiber activity elicited by the corneal air puff is relayed via the trigeminal nucleus to inferior olive projection (Freeman and Steinmetz, 2011; Ohmae and Medina, 2015; Rasmussen et al., 2008; Rasmussen et al., 2014; ten Brinke et al., 2015). We found similar climbing fiber activity during both unconditioned and conditioned eyeblink responses in the IN and FN. It might be interesting to investigate the input regions that innervate different inferior olivary regions and provide synergistic climbing fiber inputs to these two regions during DEC.
Interestingly, both the IN and FN are essential for generating conditioned eyeblink responses (Thompson, Mauk, Yeo, REFS; current study), indicating that they are not functional redundant. Importantly, the UR can only be significantly affected by photo-perturbing the FN, but not the IN. These results imply that
inhibiting FN activity specifically deregulates the output of facial motor neurons. We thus hypothesize that, whereas the burst activity of the IN-RN pathway exclusively drives the CS related blinking response, the activity of the FN pathway tonically modulates the excitability of facial motor neurons, which is critical during both conditioned and unconditioned reflexes.
Future directions for FN Output in Eyelid Closure
Our results unequivocally illustrate the DEC related modulation in the FN module. Taking into account the DEC related FN neurons predominantly increased their firing frequency during CS (Figure 1E-1I) and modulations were found in the excitatory output neurons (Figure 3D), a possible follow up will be to determine the necessity of FN neuron inhibitory outputs (Kros et al., 2015). GABA neuronal
silencing targeting the FN ipsilateral to the trained eye will potentially show conditioned and unconditioned behaviour alterations. Based on the result, a direct functional involvement of FN output in mediating both CR and UR could be established. Furthermore, exploration of the perturbation of the FN module during the animal learning process may provide additional support on the modules
involvement in CRs and URs acquisition. Another interesting question is to examine whether FN activity is essential for CR expression during the CS-US interval.
Suppression of FN neuronal activity during the CS-US interval might potentially reduce the percentage and amplitude of CRs. Taken together, the results could elucidate the functional involvement of the FN module in controlling eyelid closure, and highlight its unique role in modulating both CR and UR responses. Previous studies have unequivocally established a key cerebellar pathway which IN neurons innervate the premotor red nuclei neurons, and in turn excite 7N motor neurons (Gonzalez-Joekes and Schreurs, 2012; Morcuende et al., 2002; Sun, 2012). The next steps on the topic are to concoct both functional and anatomical studies to clarify roles of FN in associative learning and behaviour.
Figure 1. Extracellular Electrophysiological Recordings of FN Neurons during DEC
(A - C) Experimental schematics.
(A) Animals were head-fixed on a treadmill with a green LED light as the conditioned stimulus (CS) and a peri-orbital air puff as the unconditioned stimulus (US). (B) Conditioned responses (CRs, green) emerged prior to the unconditioned responses over days (URs, red) during DEC training. (C) DiI labelled track in cerebellum showing the recording location in FN.
(D) CS-related modulations of FN neurons. Top and middle rows: example eyelid and spike traces of neurons with CS-related facilitation (left), suppression (middle) and no modulation (right). Bottom row: the average neural activities of each
modulation type (From left to right, n = 86 for facilitation neurons; n = 16 for suppression neurons; n = 60 for non-modulation neurons). Green shading indicates the CS-US interval.
(E and F) Positive trial-by-trial correlation between neural facilitation and CR peak amplitude. (E) Example cell with positive correlation between neuron activity (left) and CR amplitudes (right) (Linear regression model, p < 0.001). Each row on the left panel represents a single trial of recording, ordered based on the peak amplitude of facilitation. Each row on the right panel represents the corresponding CR amplitude of the same trial. Dashed lines indicate CS and US onsets. (F) Summary of all facilitation cells with positive correlations. Red lines denote significantly correlated cells (p < 0.05, n = 9); pink lines denote nonsignificant-correlated ones (p > 0.05, n = 45).
(G and H) Same as E and F but for neurons with negative correlation between facilitation and CR peak amplitude. (G) Example of a negatively correlating FN neuron (Linear regression model, p < 0.001), trials are ordered by the facilitation amplitude. (H) Red lines denote significantly correlated cells (p < 0.05, n = 5); pink lines denote nonsignificant-correlated ones (p > 0.05, n =27).
(I) Average correlation matrix of 86 facilitation cells. Each epoch indicates the mean2 value of trial-by-trial correlation between the FN activity and eyelid closure at a given time point throughout the task. Most-correlated epochs (bright pixels) are located above the diagonal line and before the US onset. CS and US onsets are denoted with dashed line in both dimension.
(J) Summary of the onset timing of spike modulation and CR for all facilitation cells (paired t-test, p < 0.001). (K) Same as (J), but for the comparison of peak timings (paired t-test, p < 0.01).
See also Figure 3 for analysis of FN neuron responses to UR and Figure 2 for trial-by-trial correlation analysis of suppression cells
Figure 2. UR related modulation in FN neurons.
(A) Both UR related facilitation (up dark red, n = 80) and suppression (down orange, n = 6) were found in FN neurons with CR facilitation. In each panel, top row: example recording of single trial; middle row: raster plot of spike events for each representative cell; bottom row: average neural activity of each type.
(B & C) Same as (A), but for the UR related modulations in CR suppression (b; up: n = 3, down: n = 13) and no CR modulation cells (c; up: n = 31, down: n = 29).
(D) In the neurons with both CR and UR related facilitation (see a, top), the magnitudes of CR and UR related modulations are highly correlated (P = 1.4×10-13, r2 = 0.5).
(E) FN neurons with CR related modulation tend to have larger UR related modulation (left: facilitation, n = 113; right: suppression, n = 49) in FN cells with and without CR modulation (***P < 0.001, **P < 0.01, paired t-test).
Figure 3. Trial-to-trial correlation between FN neuron suppression and CR amplitude, Data related to Fig. 1
(A & B) Trial-by-trial correlation of FN neuron suppression and CR peak amplitude.
(A) Example cell with negative correlation between neuron activity (left) and CR amplitudes (right) (Linear regression model, P < 0.01). Each row on the left panel represents a single trial of recording, ordered based on the peak amplitude of suppression. Each row on the right panel represents the corresponding CR amplitude of the same trial. Dashed lines indicate CS and US onsets. (B) Summary of all suppression cells. Dark blue line denotes significantly correlated cell (P < 0.01, n = 1); light blue lines denote non significant-correlated ones (P > 0.05, n = 15).
(C) Average correlation matrix of 16 suppression cells. Each epoch indicates the mean r2 value of trial-by-trial correlation between the FN neuron activity and eyelid closure at a given time point throughout the task. All epochs are minimally-correlated (dark pixels) within the CS-US interval. CS and US onsets are denoted with dashed line in both dimensions.
(D) Summary of the onset timing of spike modulation and CR for all suppression cells (paired t-test, P < 0.01). (E) Same as (D), but for the comparison of the peak timing between suppression and CR (paired t-test, P < 0.001).
Figure 4. Task Related Modulation in Excitatory FN Neurons and the Identification of DEC Related Vermal Regions
(A) Schematics showing viral injection, optical fiber implantation and multi-channel recording in FN of VGluT2-ires-Cre or Gad2-ires-Cre mice.
(B) ChrimsonR was expressed in VGluT2 positive FN neurons (left). ChrimsonR expressing neurons show short-latency response to 595 nm light (orange shading, right). Blue dashed line indicates epoch that firing rate exceeds 3 SDs of baseline frequency within 20 ms after the light.
(C) Same as (B), but for Gad2 positive inhibitory FN neurons.
(D) Task related modulation of VGluT2 positive neurons. Neurons are categorized based on their CR related modulations. Top and middle rows: example eyelid and spike traces of individual cells. Bottom row: average firing rate of neurons with CS related facilitation (left, n = 7), suppression (middle, n = 3) and no modulation (right, n = 5).
Figure 5. In-vivo FN neuron responses to optogenetics in VGluT2-Cre and Gad2-Cre mice. Data related to Fig. 2
(A) Quantification of neurons with different responses to optogenetic stimulation in the FN of VGluT2-Cre mice with Flex-ChrimsonR injection. Summary of the proportion (up) and the numbers of FN neurons (down) in VGluT2-Cre mice showing short-latency facilitation (onset < 20 ms, dark green), long-latency facilitation (onset ≥ 20 ms, light green) and no modulation (grey).
(B) Same as (a), but in Gad2-Cre mice showing short-latency facilitation (onset < 20 ms, dark blue), suppression (light blue) and no modulation (grey).
(C) Different neuron responses to optogenetic activation in VGluT2-Cre mice. From left to right: cells responded to optical light (orange shading) with short-latency facilitation (n = 15), long-latency facilitation (n = 30) and no modulation (n = 41). Top row: example spike traces; middle row: raster plot of spike events for each representative cell; bottom row: average neural activity of each group.
(D) Same as (C), but for recordings from Gad2-Cre mice. From left to right: cells responded to optogenetics with short-latency facilitation (n = 8), suppression (n = 15) and no modulation (n = 4).
References:
Gormezano, I., Schneiderman, N., Deaux, E. & Fuentes, I. Nictitating membrane: classical conditioning and extinction in the albino rabbit. Science 138, 33-34 (1962). Patterson, M.M., Olah, J. & Clement, J. Classical nictitating membrane conditioning in the awake, normal, restrained cat. Science 196, 1124-1126 (1977).
McCormick, D.A. & Thompson, R.F. Cerebellum: essential involvement in the classically conditioned eyelid response. Science 223, 296-299 (1984).
De Zeeuw, C.I. & Yeo, C.H. Time and tide in cerebellar memory formation. Curr Opin Neurobiol 15, 667-674 (2005).
Koekkoek, S.K., et al. Cerebellar LTD and learning-dependent timing of conditioned eyelid responses. Science 301, 1736-1739 (2003).
Chen, L., Bao, S. & Thompson, R.F. Bilateral lesions of the interpositus nucleus
completely prevent eyeblink conditioning in Purkinje cell-degeneration mutant mice. Behav Neurosci 113, 204-210 (1999).
Ten Brinke, M.M., et al. Dynamic modulation of activity in cerebellar nuclei neurons during pavlovian eyeblink conditioning in mice. Elife 6 (2017).
Mauk, M.D., Li, W., Khilkevich, A. & Halverson, H. Cerebellar mechanisms of
learning and plasticity revealed by delay eyelid conditioning. International review of neurobiology 117, 21-37 (2014).
Ten Brinke, M.M., et al. Evolving Models of Pavlovian Conditioning: Cerebellar Cortical Dynamics in Awake Behaving Mice. Cell Rep 13, 1977-1988 (2015). Heiney, S.A., Wohl, M.P., Chettih, S.N., Ruffolo, L.I. & Medina, J.F. Cerebellar-dependent expression of motor learning during eyeblink conditioning in head-fixed mice. J Neurosci 34, 14845-14853 (2014).
Steinmetz, A.B. & Freeman, J.H. Localization of the cerebellar cortical zone
mediating acquisition of eyeblink conditioning in rats. Neurobiol Learn Mem 114, 148-154 (2014).
Mostofi, A., Holtzman, T., Grout, A.S., Yeo, C.H. & Edgley, S.A. Electrophysiological localization of eyeblink-related microzones in rabbit cerebellar cortex. The Journal of neuroscience : the official journal of the Society for Neuroscience 30, 8920-8934 (2010).
Van der Want, J.J., Wiklund, L., Guegan, M., Ruigrok, T. & Voogd, J. Anterograde tracing of the rat olivocerebellar system with Phaseolus vulgaris leucoagglutinin (PHA-L). Demonstration of climbing fiber collateral innervation of the cerebellar nuclei. J Comp Neurol 288, 1-18 (1989).
Heckroth, J.A. & Eisenman, L.M. Parasagittal organization of mossy fiber collaterals in the cerebellum of the mouse. J Comp Neurol 270, 385-394 (1988).
Johansson, F., Hesslow, G. & Medina, J.F. Mechanisms for motor timing in the cerebellar cortex. Curr Opin Behav Sci 8, 53-59 (2016).
Jirenhed, D.A., Rasmussen, A., Johansson, F. & Hesslow, G. Learned response sequences in cerebellar Purkinje cells. Proc Natl Acad Sci U S A 114, 6127-6132 (2017).
Ohmae, S. & Medina, J.F. Climbing fibers encode a temporal-difference prediction error during cerebellar learning in mice. Nat Neurosci 18, 1798-1803 (2015).
Carey, M.R. Synaptic mechanisms of sensorimotor learning in the cerebellum. Current opinion in neurobiology 21, 609-615 (2011).
Boele, H.J., Koekkoek, S.K., De Zeeuw, C.I. & Ruigrok, T.J. Axonal sprouting and formation of terminals in the adult cerebellum during associative motor learning. J Neurosci 33, 17897-17907 (2013).
Batton, R.R., 3rd, Jayaraman, A., Ruggiero, D. & Carpenter, M.B. Fastigial efferent projections in the monkey: an autoradiographic study. J Comp Neurol 174, 281-305 (1977).
Zhang, X.Y., Wang, J.J. & Zhu, J.N. Cerebellar fastigial nucleus: from anatomic construction to physiological functions. Cerebellum Ataxias 3, 9 (2016).
Yamada, T., et al. Sensory experience remodels genome architecture in neural circuit to drive motor learning. Nature (2019).
Gao, Z., et al. A cortico-cerebellar loop for motor planning. Nature 563, 113-116 (2018).
Gonzalez-Joekes, J. & Schreurs, B.G. Anatomical characterization of a rabbit cerebellar eyeblink premotor pathway using pseudorabies and identification of a local modulatory network in anterior interpositus. J Neurosci 32, 12472-12487 (2012). Morcuende, S., Delgado-Garcia, J.M. & Ugolini, G. Neuronal premotor networks involved in eyelid responses: retrograde transneuronal tracing with rabies virus from the orbicularis oculi muscle in the rat. J Neurosci 22, 8808-8818 (2002).
Uusisaari, M.Y. & Knopfel, T. Diversity of neuronal elements and circuitry in the cerebellar nuclei. Cerebellum 11, 420-421 (2012).
Ruigrok, T.J. Ins and outs of cerebellar modules. Cerebellum 10, 464-474 (2011). Rasmussen, A., Jirenhed, D.A. & Hesslow, G. Simple and complex spike firing
patterns in Purkinje cells during classical conditioning. Cerebellum 7, 563-566 (2008). Kros, L., et al. Cerebellar output controls generalized spike-and-wave discharge occurrence. Ann Neurol 77, 1027-1049 (2015).
Sun, L.W. Transsynaptic tracing of conditioned eyeblink circuits in the mouse cerebellum. Neuroscience 203, 122-134 (2012).
Giovannucci, A., et al. Cerebellar granule cells acquire a widespread predictive feedback signal during motor learning. Nat Neurosci 20, 727-734 (2017).
Yeo, C.H., Hardiman, M.J. & Glickstein, M. Classical conditioning of the nictitating membrane response of the rabbit. II. Lesions of the cerebellar cortex. Exp Brain Res 60, 99-113 (1985).
Yeo, C.H., Hardiman, M.J. & Glickstein, M. Classical conditioning of the nictitating membrane response of the rabbit. I. Lesions of the cerebellar nuclei. Exp Brain Res 60, 87-98 (1985).
Esposito, M.S., Capelli, P. & Arber, S. Brainstem nucleus MdV mediates skilled forelimb motor tasks. Nature 508, 351-356 (2014).
Cohen, B. Pontine reticular formation neurons and motor activity. Science 199, 207-208 (1978).
Siegel, J.M. & McGinty, D.J. Pontine reticular formation neurons: relationship of discharge to motor activity. Science 196, 678-680 (1977).
Takakusaki, K., Chiba, R., Nozu, T. & Okumura, T. Brainstem control of locomotion and muscle tone with special reference to the role of the mesopontine tegmentum and medullary reticulospinal systems. J Neural Transm (Vienna) 123, 695-729 (2016). Gao, Z., van Beugen, B.J. & De Zeeuw, C.I. Distributed synergistic plasticity and cerebellar learning. Nature reviews. Neuroscience 13, 619-635 (2012).
Najac, M. & Raman, I.M. Synaptic excitation by climbing fibre collaterals in the cerebellar nuclei of juvenile and adult mice. J Physiol 595, 6703-6718 (2017). Medina, J.F., Nores, W.L., Ohyama, T. & Mauk, M.D. Mechanisms of cerebellar learning suggested by eyelid conditioning. Curr Opin Neurobiol 10, 717-724 (2000).
Na, J., Sugihara, I. & Shinoda, Y. The entire trajectories of single pontocerebellar axons and their lobular and longitudinal terminal distribution patterns in multiple aldolase C-positive compartments of the rat cerebellar cortex. J Comp Neurol (2019). Biswas, M.S., Luo, Y., Sarpong, G.A. & Sugihara, I. Divergent projections of single pontocerebellar axons to multiple cerebellar lobules in the mouse. J Comp Neurol (2019).
Gao, Z., et al. Excitatory Cerebellar Nucleocortical Circuit Provides Internal Amplification during Associative Conditioning. Neuron 89, 645-657 (2016). Freeman, J.H. & Steinmetz, A.B. Neural circuitry and plasticity mechanisms underlying delay eyeblink conditioning. Learn Mem 18, 666-677 (2011).
Rasmussen, A., Jirenhed, D.A., Wetmore, D.Z. & Hesslow, G. Changes in complex spike activity during classical conditioning. Front Neural Circuits 8, 90 (2014).
