As far as horizontal eye movements are concerned, the well-known hypothesis, not yet experimentally proved, of the common neural integrator states that the eye-position signal is generated by a common network, regardless of the type of versional movement. The aim of this study was to evaluate the validity of this hypothesis by checking whether the sensitivity to eye position of the neurones of the nucleus prepositus hypoglossi (NPH) (the main component of the system integrating the different incoming velocity signals) would be the same regardless of the type of versional movement. 2. The discharge of sixty-five NPH neurones was recorded in the alert cat during spontaneous eye movements made in the light and in response to sinusoidal rotations of the head in complete darkness. 3. For each NPH neurone, the sensitivity to eye position was determined from measurements carried out during intersaccadic fixation. The discharge rate of the studied neurone was plotted against eye position. The slope of the resulting regression line gave the sensitivity (measured during intersaccadic fixation in the light) of the neurone to eye position, which was termed K(f). 4. A new method was developed to measure the sensitivity to eye position (K(v)) of neurones during vestibular slow phases. The difficulty came from the fact that, during slow phases, eye velocity and eye position changed simultaneously and that each of those two variables could influence neuronal activity. For each neurone, the instantaneous firing rate was measured each time the eye passed through a given position during any slow phase generated during any vestibulo-ocular reflex. At a given position, the discharge rate of the neurone under study was plotted against the eye velocity. From the resulting linear regression line, two interesting values were obtained: its slope, corresponding to the sensitivity of the neurone to eye velocity, R(v), (at that given eye position) and its 'y'-intercept, F(0), the interpolated firing rate when the eye velocity was zero. This procedure was repeated for different eye positions. The values of F(0) were then plotted against the eye positions. The slope of the resulting regression line gave the sensitivity (measured during vestibular stimulation) of the neurone to eye position, which was termed K(v). 5. The errors on the individual values of K(f) and K(v) were assessed in order to allow a statistical comparison at the single unit level. 6. We found that, for each of our sixty-five neurones, the sensitivity to eye position measured during intersaccadic fixation in the light was equal to the sensitivity to eye position measured during the vestibulo-ocular reflex (VOR) elicited in complete darkness. We conclude that our results favour the hypothesis of a unique horizontal oculomotor integrator for all versional movements.
