Reconciling functional differences in populations of neurons recorded with two-photon imaging and electrophysiology
- MindScope Program, Allen Institute, United States
- Allen Institute for Brain Science, Allen Institute, United States
Figures
Figure 1
Overview of the ephys and imaging datasets.
(A) Illustration of the orthogonal spatial sampling profiles of the two modalities. Black and white area represents a typical imaging plane (at approximately 250 µm below the brain surface), while …
Figure 2 with 4 supplements
Baseline metric comparison.
(A) Steps involved in computing response magnitudes for units in the ephys dataset. (B) Same as A, but for the imaging dataset. (C) Drifting gratings spike rasters for an example ephys neuron. Each …
Figure 2—figure supplement 1
Baseline comparisons for additional preference and selectivity metrics.
(A) Distribution of preferred directions across five areas. (B) Distribution of preferred orientations across five areas. (C) Distribution of preferred spatial frequency across five areas. (D) …
Figure 2—figure supplement 2
Matching laminar distribution patterns across modalities.
(A) Original layer distributions, prior to resampling, for imaging (green) and ephys (gray). (B) Layer distributions following the resampling procedure, for imaging (light green) and ephys (light …
Figure 2—figure supplement 3
Characterizing the impact of running behavior on response metrics.
(A) Average speed and running fraction for individual mice from imaging and ephys experiments, grouped by genotype. Shaded areas represent the mean ± standard deviation for each genotype. Colors are …
Figure 2—figure supplement 4
Matching running behavior across modalities.
(A) Cumulative histogram of drifting gratings running fractions for the ephys dataset, and for the imaging dataset before and after performing the behavioral matching procedure. (B) Fraction of …
Figure 3
Impact of calcium event detection on functional metrics.
(A) GCaMP6f fluorescence trace, spike times, and detected events for one example neuron with simultaneous imaging and electrophysiology. Events are either detected via exact ℓ0-regularized …
Figure 4
Comparing responses across GCaMP-expressing mouse lines.
(A) GCaMP expression patterns for the four lines used for ephys experiments, based on two-photon serial tomography (TissueCyte) sections from mice used in the imaging pipeline. (B) Unit yield …
Figure 5 with 2 supplements
Effects of the spikes-to-calcium forward model.
(A) Example of the forward model transformation. A series of spike times (top) is converted to a simulated ΔF/F trace (middle), from which ℓ0-regularized events are extracted using the same method …
Figure 5—figure supplement 1
Sampling the entire space of forward model parameters.
(A) Schematic of three different ‘slices’ through forward model parameter space, sampling all values of unitary amplitude (A), decay time (τ), or noise level (σ). (B) Fraction of neurons responding …
Figure 5—figure supplement 2
Relationship between forward model parameters, genotype, and functional metrics.
(A) Distribution of noise level (σ), decay time (τ), and amplitude (A) parameters for the 10 different Cre lines used in this study. (B) Effect of sub-selecting neurons based on drifting gratings …
Figure 6 with 1 supplement
Sub-sampling based on event rate or ISI violations.
(A) Top: Change in fraction of neurons responding to drifting gratings for each area in the imaging dataset, as function of the percent of imaged neurons included in the comparison. Middle: …
Figure 6—figure supplement 1
An alternative metric of response robustness.
(A) Relationship between response reliability (fraction of preferred-condition trials with responses larger than 95% of baseline intervals) and coefficient of variation (ratio between standard …
Figure 7
Summary of results for all stimulus types.
(A) Jensen–Shannon distance between the ephys and imaging distributions for three metrics and four stimulus classes. Individual dots represent visual areas, and horizontal lines represent the mean …
Figure 8
Identifying functional classes based on response reliability.
(A) Example clustering run for all neurons in the ephys dataset. First, unsupervised clustering is performed on the matrix of response reliabilities to four stimulus types. Colors on the t-SNE plot …
Author response image 1
