Axial responses in visual cortical cells: Spatio-temporal mechanisms quantified by Fourier components of cortical tuning curves

Abstract

The responses of 81 cells from area 17 in paralysed and aneasthetized cats were studied with moving spots and moving bars of different lengths. Tuning curves were measured and plotted as polar-plots. The strongest response of visual cortical cells to a moving bar occurs when the stimulus trajectory crosses the long axis of the receptive field (Hubel and Wiesel 1962). The optimal orientation for a moving and a flashing bar are identical, so that this response-type has been called the orientational component. For a moving spot, however, in most cases the strongest response occurs for motion along the receptive field long axis (axial component). Thus, the axial and orientational components are orthogonal (Wörgötter and Eysel 1989). It is shown that orientational and axial components can display direction selectivity and for short bar stimuli a superposition of the two orthogonal components is demonstrated. Such a superposition in general, resulted in a polar-plot with four peaks 90° apart from each other (four-symmetrical polar-plot). Polar-plots with three or two response peaks were also found; the actual number of response peaks depending on the direction selectivity of the components. In many cells pure axial responses could be elicited with a light spot which stimulates only motion dependent mechanisms. Thus, it was concluded that temporal facilitation is strongly involved in the generation of axial responses. Fourier analysis of polar-plots (SDO-analysis, Wörgötter and Eysel 1987; Wörgötter et al. 1990) was applied to determine the tuning strengths of the different components. In correspondence with the periodicities of a moving oriented stimulus in the visual field, the first harmonic represents directional selectivity, the second orientation selectivity and the fourth harmonic was used to quantify the four-symmetrical superposition effect. It was statistically shown that the strongest superposition (i.e. largest fourth harmonic) occurred for intermediate bar lengths (1–2°). For longer bars only the orientational and for shorter bars predominantly the axial component occurred. In monkey visual cortex (V1, V3) four-symmetrical polar-plots can be obtained even with long stimuli (De Valois et al. 1982; Felleman and Van Essen 1987). Consequentially, we show that a strong fourth order Fourier component occurs. This supports the importance of quantification of higher order symmetries in cortical tuning curves by higher order harmonics in SDO-analysis.