Fig. 7: Intermediate spatial model analysis along the T-L Parameter Continuum and F-L Coordinate Continuum for FEF.
From: Integration of landmark and saccade target signals in macaque frontal cortex visual responses
a Intermediate spatial models for both (1) Target-Landmark coding in the same fixation-centered coordinate system (the T-L Parameter Continuum) and (2) target coding between fixation-centered and landmark-centered coordinates (F-L Coordinate Continuum). Black circles show intermediate steps along the mathematical continuum between TF(e) (represented as 0) and one of the other two models (represented as 1.0). b Example distribution of mean residuals of fit along the F-L Coordinate Continuum for the neuron shown in Fig. 4, displaying minimal residuals (best fit) exactly at TF(e). c–h Complete Continuum analysis for all FEF neurons (see Supplementary Fig. 5 for a similar SEF analysis). In each panel, fits to experimental data are plotted as a function of the median best fit for their corresponding 100 shuffled control datasets (i.e., along x-axis each dot represents median of 100 shuffles) and their overall distributions are shown above/beside the plot. The magenta lines show the intersection of medians along x and y axes and the diagonal dotted lines show the line of equality between the pooled and shuffled control fits. To be comprehensive, we included neurons with and without spatial tuning in this analysis. c–e Swarm plot charts of the distributions of best fits along the T-L Parameter continuum. c Fits for response fields where data were pooled across different target-landmark configurations. d Configuration-dependent analysis. Each color represents a specific target-landmark configuration. e Recombined data (mean of 4 median fits/cell). f–h Similar analysis (as c–e) but for the F-L Coordinate Continuum. f Fits for pooled configuration data. g Configuration-dependent Fits. h Recombined data fits. Note that FEF/SEF neurons that were spatially tuned always showed a significantly shifted coding toward both LF(e) and TL(e) for at least one target-landmark configuration, whereas untuned neurons never showed significant shifts along either continuum for any TLC. The colorized dots in (c) and (f) correspond to TF(e) (green) and LF(e) (blue) preferring neurons shown in Fig. 6.
