Purchase this article with an account.
Russell Hamer, Givago Souza, Thiago Costa, Bruno Gomes, Luiz Carlos Silveira, Dora Ventura; Principal components analysis (PCA) of human cortical contrast response functions reveals two dominant mechanisms: M and P?. Invest. Ophthalmol. Vis. Sci. 2013;54(15):3409.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
VEP contrast response functions (CRFs) often manifest two linear limbs on log-contrast (C) axes - a shallow limb at low C, and a steep limb at high-C. These have been hypothesized to reflect M- and P-dominated responses, respectively (e.g., Nakayama & Mackeben, 1982; Bobak et al., 1984; Tyler & Apkarian, 1985; Souza et al., 2007, 2008). We have applied the agnostic procedure of PCA analysis to cortical CRFs obtained with the sweep VEP (sVEP) over a range of spatio-temporal S-T conditions.
sVEPs were recorded from 11 subjects while presenting 0.4, 2, 10 or 16 c/d achromatic sinewave gratings (161 cd/m2 mean lum) phase-reversing at 4.29, 6, 10 or 15 Hz. Contrast increased logarithmically from 1.4 to 90% during each 10-s trial. A vector average response at 2F was calculated (10 sweeps/condition). PCA analysis was applied to the sVEP amplitudes from 153 CRFs.
Two PCA components explained most of the variance (PC1, PC2). All CRFs were reconstructed based on PC1 and PC2. For 2-limbed CRFs, PC1 dominated at low C and significant PC2 amplitude began only for C > 10 - 20%. For 1-limbed CRFs, only PC1 was significant. Both PCs were ~linear on log-C axes. A Naka-Rushton function was fit to each PC, on linear C axes, for all S-T combinations, providing an estimate of contrast gain (Cgain = Rmax/Csat50). Cgains for PC1 were greater than PC2 gains by 1 to 1.5 log units at all S-T combinations. The highest Cgains for PC1 occurred at TFs ≤10 Hz and were ~equal across SF up to 10 c/deg. The lowest gains tended to occur at 15 Hz across SF. For PC2, the highest Cgains occurred at intermediate to high S-T combinations. Analysis of individual 2-limbed CRFs showed that the C at which PC2 starts to contribute significantly is the C at which the sVEP amplitude transitions from one slope to the other and where the phase response suddenly changes (Fig. 1).
PCA analysis of sVEP CRFs provides support for the working hypothesis that M and P pathway signals combine to generate the overall sVEP CRF, and that these are, to a large degree, separable, with PC1 and PC2 being associated with M and P, respectively.
This PDF is available to Subscribers Only