Fig 2. Calculation of dimensionality and percent shared variance.

(A) Factor analysis partitions the spike count covariance of sampled excitatory and inhibitory neurons together into shared and independent components. (B) Shared dimensionality (d_shared) is the minimum number of eigenvectors of the shared variance matrix necessary to explain 95% of shared variance. Modes are sorted by shared variance explained along the x-axis. (C) Percent shared variance is the ratio of shared to total variance (i.e., shared / (shared + independent)). The percent shared variance is first computed for each neuron, then averaged across all neurons of the same type.