Abstract
Abstract: :
Purpose: Luminance, color, orientation, spatial frequency, motion and stereo disparity are local features that help segregate an object from its background. The segmentation of higher visual processes can be revealed electrophysiologically with the tsVEP. The present study aimed at characterizing, in terms of amplitude and latency, the ts peaks defined by motion or orientation, and at assessing the upper limit of the tsVEP's temporal-frequency transfer function. Method: We recorded VEPs in 10 normal adult subjects with visual acuities of 20/20 or better. For the first experiment, two stimuli were used: (1) oriented line segments (0.1° width, 90% contrast), in parallel or changing in orientation, giving the appearance of a checkeboard; (2) random dot patterns (30% contrast), where all dots moved in the same direction or oppositely, resulting in motion-defined checkerboard. The ts curve was derived by substracting the mean response obtained with homogenous stimuli (all oriented in parallel or moving in the same direction) from the mean response obtained with textured stimuli (arrangement of line segments or moving squares that give the impression of a checkerboard). This linear combination has been shown to cancel out all "low level" VEP (llVEP) components, resulting in the "tsVEP". In the second experiment, the orientation-defined stimuli alternated between the homogeneous and texture conditions at different temporal frequencies (between 7.6 and 45 Hz). tsVEP components were isolated from the llVEP in the frequency domain. Results: For orientation, the tsVEP appeared as a negative peak of 12uV at around 150ms; clearly later than any llVEP, and earlier than typical cognitive responses. This peak was more sharply defined than when ts was based on motion gradients, where a double peak appeared. In the second experiment the temporal tuning had a bell-shaped characteristic, peaking at 9 Hz for the tsVEP and at 11.2 Hz for the llVEP. The tsVEP declined earlier than the llVEP and appeared shifted on the temporal frequency axis by a factor of two. Conclusion: Gradients in orientation and motion elicit a negative VEP component around 150ms, which has further substructure for the motion dimension. Also, the more rapidly declining tsVEP compared to llVEP as temporal frequency increases suggests that tsVEP requires more complex processing of the information than llVEP, supporting one of two recent contradictory psychophysical studies of this phenomenon.
Keywords: 394 electrophysiology: non-clinical • 597 texture • 621 visual cortex