The functional equivalence of ascending and parallel fiber inputs in cerebellar computation

Abstract

At the center of the computational cerebellar circuitry are Purkinje cells, which integrate synaptic inputs from >150,000 granule cell inputs. Traditional theories of cerebellar function assume that all granule cell inputs are comparable. However, it has recently been suggested that the two anatomically distinct granule cell inputs, ascending and parallel fiber, have different functional roles. By systematically examining the efficacy of patches of granule cells with photostimulation, we found no differences in the efficacy of the two inputs in driving the activity of, or in producing postsynaptic currents in, Purkinje cells in cerebellar slices in vitro. We also found that the activity of Purkinje cells was significantly increased upon stimulation of lateral granule cells in vivo. Moreover, when we estimated parallel fiber and ascending apparent unitary EPSC amplitudes using photostimulation in cerebellar slices in vitro, we found them to be indistinguishable. These results are inconsistent with differential functional roles for these two inputs. Instead, our data support theories of cerebellar computation that consider granule cell inputs to be functionally comparable.

Figure 1.

Mapping the efficacy of patches of granule cells in driving the firing of Purkinje cells. A, Glutamate-evoked currents were recorded in a voltage-clamped Purkinje cell (−60 mV) in response to photolysis of glutamate over its dendritic tree in the molecular layer (top) and in the granule cell layer beneath its soma (bottom). The kinetics of the currents resulting from photorelease of glutamate in the granule cell layer were much slower and the noise of the currents was much greater than that observed when glutamate was photoreleased directly over the dendrites of the target Purkinje cell. B, To examine the spatial selectivity of the 40 μm spot used for photolysis of glutamate in the X–Y plane, the dendrites of voltage-clamped Purkinje cells in acutely prepared coronal slices of the cerebellum were used as glutamate detectors. Using the same intensity pulse of UV light, glutamate was photoreleased at various lateral distances from the center of the dendritic tree of a voltage-clamped Purkinje cell, and the resulting currents were recorded. The photolysis spot was moved laterally in 10 μm increments in both directions. Sample glutamate-evoked currents recorded from a single Purkinje cell in response to glutamate photolysis at various lateral distances from the center of its dendritic tree are shown together with an average plot of peak EPSC amplitude as a function of photolysis location (n = 3 cells). Each raw data trace is the average of five trials. C, Experiments were performed to measure the depth of photolysis in the Z plane. Purkinje cells on the surface of sagittal cerebellar slices were voltage clamped and the intensity of the 40 μm spot of UV light was adjusted to produce a large-amplitude inward current when positioned on the surface of the slice within dendritic tree of the target Purkinje cell (0 μm from dendrite). The plane of focus was then moved vertically above the slice in 20 μm increments and the resulting currents in response to the same pulse of UV light at each location recorded. Sample glutamate- evoked currents recorded from a single Purkinje cell in response to glutamate photolysis at various vertical distances from its dendritic tree are shown together with an average plot of peak EPSC amplitude as a function of vertical distance from the Purkinje cell dendrite (n = 4 cells). Each raw data trace is the average of five trials. D, Glutamate photoreleased by a constant-intensity UV pulse activated patches of granule cells and the firing of a single Purkinje cell in a coronal cerebellar slice was monitored with extracellular recordings. Lower traces show instantaneous firing rate in response to photolysis of glutamate using the same intensity pulse of UV light at two separate locations. ML, Molecular layer; PCL, Purkinje cell layer; GCL, granule cell layer; WM, white matter. E, Top, Schematic drawing of the cerebellar slice and the Purkinje cell whose activity was recorded (shown in red). The grayscale map shows the maximum poststimulus instantaneous firing rate of the target Purkinje cell after photolysis of glutamate at each location. The instantaneous firing rate response at each location is superimposed on each location. The baseline firing rate of this Purkinje cell was 50 spikes/s. Bottom, The instantaneous maximum firing rates obtained from photolysis at each location within a column were averaged and these column strengths were then normalized to the strongest column value. F, The normalized column strengths obtained from eight maps were averaged based on lateral distances from the soma of the target Purkinje cell and renormalized to the strongest column value. Normalized strengths predicted for patches of granule cells making only parallel fiber inputs based on various estimates of granule cell connection probability and EPSC amplitudes are also shown as blue and red bars (see Results for details). Results are shown as mean ± SEM for these and all subsequent figures. * denotes statistical significance with respect to the column denoted by the arrow at p < 0.01. G, For all eight experiments included in F above, two investigators independently determined the location of the “ascending” patch in the row closest to the Purkinje cell layer. The average maximum instantaneous firing rate at the location that corresponded to the position where granule cells form ascending inputs (Asc patch) was taken as one. The relative firing rates of the adjoining patches located on either side of the ascending patch and the second adjacent pair located 80 μm away from the ascending patch were pooled and averaged. Data are presented as mean ± SEM.

Figure 2.

High-resolution EPSC maps in juvenile rats. Top, A sample experiment in which EPSCs were monitored in a whole-cell voltage-clamped Purkinje cell in response to glutamate photorelease as in Figure 1E. In these experiments the spot was moved in 20 μm increments to improve mapping resolution. Inclusion of 100 μm Alexa Fluor 488 in the patch-pipette solution allowed determination of the exact orientation of the dendritic plane of the voltage-clamped Purkinje cells (red). The grayscale patches show the average magnitude of peak EPSC amplitudes resulting from glutamate photolysis at each location. The average of three EPSCs (red traces) is also shown for each position. Bottom, EPSC peak amplitudes, 20–80% rise times, and 20–80% decay times from every position in a column were averaged and normalized to the largest column value.

Figure 3.

High-resolution mapping confirms the relative homogeneity of the granule cell population. A, The average normalized EPSC map and dendritic orientation (red) obtained from six Purkinje cells from experiments similar to that presented in Figure 2. B, A row-by-row analysis of the average normalized EPSC map shown in A. Each row was normalized to the largest value in that row. C, A column-by-column analysis of the average normalized EPSC map shown in A. Shown is the normalized average of the seven rows shown in A. * denotes statistical significance with respect to the column denoted by the arrow at p < 0.005. D, For each of the EPSC mapping experiments averaged in A above, two investigators independently determined the location of the “ascending” patch in the row closest to the Purkinje cell layer. The average peak EPSC amplitude of the position that corresponded to the location where granule cells form ascending synapses (Asc patch) was taken as one. The relative peak EPSC amplitudes of the adjoining patches located on either side of the ascending patch and the second adjacent pair located 40 μm away from the ascending patch were pooled and averaged. Data are presented as mean ± SEM.

Figure 4.

Adult maps also do not show powerful ascending granule cell patches. A, A single EPSC mapping experiment as shown in Figure 2 except in a slice obtained from an adult rat (2–3 months old). The dendritic plane of the voltage-clamped Purkinje cell is shown in red. The grayscale map shows the average magnitude of peak EPSC amplitudes resulting from glutamate photolysis at each location. Average of three EPSCs (red traces) is also shown for each position. Bottom, Peak EPSC amplitudes from every position in a column were averaged and normalized to the largest column value. B, Top, Average normalized EPSC map and dendritic orientation (red) obtained from five Purkinje cells in slices from adult rats. Bottom, A column-by-column analysis of the average normalized EPSC map shown above. Peak EPSC amplitudes from every position in a column were averaged and normalized to the largest column value. * denotes statistical significance with respect to the column denoted by the arrow at p < 0.05. C, Same as Figure 3D except from experiments performed in slices obtained from adult rats.

Figure 5.

The activity of Purkinje cells in vivo can be driven by granule cells which only form parallel fibers. A, Extracellular recordings were made from Purkinje cells in vivo, and patches of granule cells located 300 or 600 μm away along the parallel fiber beam were electrically stimulated using a train of three pulses at 100 Hz. B, Raw data from a representative experiment illustrating the occurrence of complex spikes (asterisk) during recording from a Purkinje cell. The pause in simple spike firing following the complex spike as denoted by asterisks in the raw data trace is shown with an expanded time scale. C, Raster plot and the corresponding poststimulus time histogram (5 ms bins) showing the response of a Purkinje cell to the repeated activation of a patch of granule cells located 300 μm away along the parallel fiber beam. D, Same as C except granule cells located 600 μm away were activated.

Figure 6.

Apparent unitary EPSC amplitudes of parallel fiber and ascending inputs obtained with photolytic activation of granule cells are indistinguishable. A, Glutamate was photoreleased in the granule cell layer within the dendritic plane of the target Purkinje cell to activate ascending inputs and in the granule cell layer at least 100 μm lateral to the dendritic plane to activate parallel fiber inputs. B, Left, The top trace shows the response of a Purkinje cell to photolysis of glutamate onto granule cells which formed parallel fiber synapses. The absolute apparent unitary EPSC amplitude obtained from the sum-of-Gaussians fit to the amplitude histogram (R² = 0.91, p = 0.9) was 15.5 pA. The inset shows the Gaussian fit to the baseline noise histogram. Right, The amplitude distribution shown in B was force fit using the sum of three Gaussians (1 for failures and 2 for responses, right panel) resulting in an absolute apparent unitary EPSC amplitude of 7.7 pA. When fitting the distribution using the sum of three Gaussians, R² = 0.86 and p = 0.2. This suggests that 15.5 pA, and not 7.7 pA, is the amplitude that represents the apparent unitary response. C, As in B except the photolysis spot was positioned to activate granule cells giving rise to ascending inputs. The absolute apparent unitary EPSC amplitude obtained from the sum-of-Gaussians fit (R² = 0.87,p = 0.9) to the amplitude histogram was 14.6 pA. D, Scatter plot of individual (open) and mean (filled) absolute apparent unitary granule cell EPSC amplitudes obtained with photolytic stimulation of the granule cell layer. The average absolute apparent unitary EPSC amplitude obtained with photolytic activation of granule cells giving rise to parallel fiber inputs (PFI) was 12.3 ± 2.3 pA (n = 6) and that giving rise to ascending inputs (AI) was 9.6 ± 1.2 pA (n = 13).