Single pulse and repetitive transcranial magnetic stimulation (TMS) over the prefrontal cortex (PFC) have been shown to interfere with memory-guided saccades when applied midway through the delay period. In this study we demonstrate disruption of memory-guided saccades when single pulse TMS is applied over the PFC coincident with target onset and describe the effect of altered TMS parameters.

With local ethical approval, 13 healthy volunteers (ages 21 - 49, 9 male) took part in a series of studies (6 subjects per study). While fixating a central light emitting diode (LED) embedded in a vertical screen, a target LED flashed on for 50 ms at one of 8 positions (±3.75, 7.5, 11.25 and 15 deg along the horizontal). The subject was required to make a saccadic eye movement to the remembered target location after a delay of 2 s, cued by central LED offset. Eye movements were recorded using an infrared eye movement detector (Skalar). TMS was delivered at 80% of the maximal stimulator intensity using a MagStim 200 stimulator connected to a 7 cm figure-of-eight coil (maximum output 2.2 Tesla). TMS parameters were varied to investigate the most effective:

1. Time of TMS, relative to target onset;
2. Position of TMS coil, on a grid of 9 points over the right prefrontal cortex;
3. Direction of induced current, out of eight directions, 45° apart.

Disturbance in memory-guided saccade performance was seen as an increase in the variable error of the final eye position of the saccade (variance of final eye position around mean final eye position). Variable error increased when TMS was applied coincident with target onset, 6 cm anterior and 2 to 4 cm lateral to the vertex. TMS mid-delay did not consistently result in a decrease in accuracy at any of the prefrontal positions. An antero-lateral induced current direction was optimal for this increase in variance. The presence of a TMS induced blink did not affect saccade accuracy. Using the same parameters, TMS did not affect reflexive saccades, suggesting that the increase in variable error was not due to a disruption of target perception. This leads us to consider that TMS applied over the PFC resulted in a disturbance in the registration phase for spatial memorisation.

In summary, single pulse TMS of the prefrontal cortex coincident with target onset was found to disturb the accuracy of memory-guided saccades. These results suggest that the PFC is involved very early in a memory-guided saccade task, which has not been previously shown with TMS. This agrees with single cell electrophysiological studies in the monkey which show phasic visual responses in the PFC to target onset of a memory-guided saccade (Funahasi et al, 1990, Journal of Neurophysiology, 63: 814-831).