Given the measured aspect ratio of the simple cell subfields (Jones and Palmer, 1987a,b) and the cross correlation studies suggesting that the subfields arise from geniculate input (Tanaka, 1983, Reid and Alonso, 1995), it seemed likely that the relay cell input to a simple cell would by itself show significant orientation selectivity. To test this conjecture, in both of the cortical inactivation experiments (Chung and Ferster, 1998, Ferster et al., 1996), the orientation selectivity of the membrane potential fluctuations evoked by visual stimulation was measured before and during cortical inactivation. Neither method of cortical inactivation significantly altered the width of orientation tuning of the visually evoked EPSPs. [Ferster et al. (1996) used drifting sinusoidal gratings as the stimulus, and only the tuning of the amplitude of temporal modulations of the voltage was measured; the mean voltage and its tuning were not measured.] Thus, the geniculate input is well tuned for orientation, as is predicted by its spatial organization (Tanaka, 1983, Reid and Alonso, 1995).
Note that these experiments also indicate that the orientation tuning of the inactivated intracortical input to these cells is very similar to that of the thalamic input. This point lends additional support to the feedforward model. The orientation tuning of the LGN input to a simple cell is not fixed; rather, it depends on the stimuli used. For example, gratings of increasing spatial frequency yield increasingly narrowly tuned LGN input (Troyer et al., 1998). For the LGN input and the cortically induced input to have identical orientation tuning, the cortical input must follow the tuning of the LGN inputs, supporting the primacy of the LGN input in determining orientation tuning.