Fig. 6. Single-trial analysis, linear regression, and decoders reveal that initial conditions predict RT but not choice.

LFADS⁴⁴ trajectories in the space of the first three orthogonalized factors (PC_1,2,3), obtained via PCA on LFADS latents, plotted for (a) the fastest 30% of trials (blue) and the slowest 30% of trials (red) for left reaches and (b) for left (purple) and right (green) reaches, all for the easiest coherence from a single session (23 units). Each trajectory is plotted from 200 ms before checkerboard onset (dots) to movement onset (diamonds). c, d Variance explained (R²)/decoding accuracy from linear/logistic regressions of binned spiking activity (20 ms) and coherence to predict trial-matched RTs/eventual choice from all 23 units in the LFADS session shown in (a and b). Horizontal black dotted lines are the variance explained by a regression using stimulus coherence to predict RTs (12%) and 50% accuracy respectively. The magenta and light green dotted lines are the 99th and 1st percentiles of R²/accuracy values calculated from an analysis of trial-shuffled spiking activity (500 repetitions) and RTs/choice. e/f R²/accuracy values, calculated as in (c) and (d), averaged across 51 sessions. Orange shaded area is SEM. e 6.32% is the average percentage of variance explained across the 51 sessions for both monkeys by regressions using stimulus coherence to predict RTs. The horizontal black dotted line in (f) denotes 50% accuracy. Blue highlight boxes in (c) and (e) denote prestimulus neural covariability with RT. Cue - Checkerboard onset. Source data are provided as a Source Data file.