Frontal non-invasive neurostimulation modulates antisaccade preparation in non-human primates

Abstract

A combination of oculometric measurements, invasive electrophysiological recordings and microstimulation have proven instrumental to study the role of the Frontal Eye Field (FEF) in saccadic activity. We hereby gauged the ability of a non-invasive neurostimulation technology, Transcranial Magnetic Stimulation (TMS), to causally interfere with frontal activity in two macaque rhesus monkeys trained to perform a saccadic antisaccade task. We show that online single pulse TMS significantly modulated antisaccade latencies. Such effects proved dependent on TMS site (effects on FEF but not on an actively stimulated control site), TMS modality (present under active but not sham TMS on the FEF area), TMS intensity (intensities of at least 40% of the TMS machine maximal output required), TMS timing (more robust for pulses delivered at 150 ms than at 100 post target onset) and visual hemifield (relative latency decreases mainly for ipsilateral AS). Our results demonstrate the feasibility of using TMS to causally modulate antisaccade-associated computations in the non-human primate brain and support the use of this approach in monkeys to study brain function and its non-invasive neuromodulation for exploratory and therapeutic purposes.

Figure 1. Behavioral paradigm illustrating the experimental antisaccade paradigm.

(Upper panel) Antisaccade paradigm practiced by the two monkeys under the online impact of sham (left panel) or active (right panel) TMS single pulses. After fixating on a central stimulus (red), monkeys were to initiate a fast saccade to a location in the opposite direction with respect to a peripheral target (green) appearing on the screen, simultaneously (no gap) to the disappearance of the central fixation. Animals performed within each block, no-TMS trials (white small rectangles) yielding no stimulation at all (Upper Left) and TMS trials (grey small rectangles) during which a single TMS pulse was delivered at a given postarget onset SOA prior to the AS initiation, to modulate the planning of visually guided oculomotor activity (Bottom panel) Example of an experimental session. Animals performed a total of 4 blocks of AS training per session. In one of the blocks they did not receive TMS (white long rectangle), whereas in the remaining 3, they received in half of the trials TMS pulses (see long grey-filled rectangles) at one of the 3 intensities used in the study (30%, 40% and 50%). The order of the four blocks (3 TMS blocks at 30%, 40% or 50% absolute TMS intensities and 1 noTMS block) was randomized within each session. Monkeys performed 100 trials per block (50 no-TMS and 50 TMS trials) for a total of 400 trials per session, and received 50 pulses per TMS block (i.e., only in 50% of the trials), amounting to 150 pulses per experimental session. Independent sessions comprising active TMS pulses delivered at 100 ms or 150 ms SOA post target onset on the FEF, sham TMS pulses and active TMS stimulation in a control location were carried over.

Figure 2. Schematic of TMS sites.

Modified picture showing a top view of each of the two monkey’s scalp profiles (animals ‘Y’ and ‘C’), while posted and under training. The dotted line corresponds to the stereotaxic zero bar; the grey dot signals the location and size of the head-post; the orange dot corresponds to the location where digit movements were evoked by TMS pulses; the red dot FEF region of stimulation; the double white/grey dots is an approximate schematic representation of the TMS figure-of-eight coil which was located on the FEF region.

Figure 3. TMS coil positioning.

Schematic drawing of the smallest of the commercially available coils, which was used for this experiment (Upper panel), a custom-made ∼25 mm loop radius figure-of-eight TMS coil (exact dimensions of the coil used indicated in the figure) (Magstim Company, Carmathenshire, Wales). (Bottom panel) X-rays photography of monkey ‘C’. The red target represents the estimated stereotaxic coordinates of the monkey’s FEF area. The length of the white bar illustrates approximate differences in bone thickness between the human and the macaque skull at frontal locations.

Figure 4. Estimate of discomfort induced by TMS.

Estimation of discomfort based on the interaction between percent of initiated trials (used as an indirect correlate of the level of discomfort; the higher the discomfort the lower the number of initiated trials) and TMS intensity (% of machine maximal output). (Upper panel) Note that below 50% intensity, low discomfort is inferred from the high percent of initiated trials (grey and black lines) respectively for monkey ‘C’ and ‘Y’, for the TMS condition (dotted lines) as compared to the sham TMS condition (solid lines). (Bottom panel) Representative traces of eye movement with/without TMS (respectively grey, black lines) at 50% intensity for one of the monkeys. At least for the two SOAs used in the study (100 and 150 ms) eye movement metrics were not affected by TMS. Furthermore, no saccades were elicited by the stimulation.

Figure 5. Saccade latencies in TMS or no-TMS trials.

Relative modulation of antisaccades latencies (mean and SD) under the impact of online FEF TMS normalized by the effects of sham TMS (white columns; (real TMS-noTMS)-(sham TMS-noTMS)) on the FEF, or active stimulation on a control cortical site (black columns; (real TMS-noTMS)-(active control TMS-noTMS)). Data are shown in millisecond differences for each of the two monkeys (‘C’ and ‘Y’) with TMS delivered at a SOA of 150 ms pre-target onset and at high intensity (50%), at which the effects of active FEF TMS were mostly noted in both monkeys (see Supplementary Table S1 for details). Decreases in normalized AS latency differences suggest a TMS-induced acceleration of AS preparation time with regards to the observed effects for sham TMS (white columns) or active TMS (black columns) in a control site and vice-versa. Notice that in both animals (‘C’ and ‘Y’) active TMS pulses decreased the average latency differences of the AS towards the hemifield ipsilateral to the stimulated FEF, whereas changes were marginal or null for contralateral AS.