Figure 1.
Proportion of phosphene detection for different levels of TMS intensity using the method of constant stimuli for 10 participants.
The data were fitted with a cumulative Weibull psychometric function to obtain 60% and 50% phosphene thresholds. Stimulation intensity refers to the percentage of stimulator's maximum output. Error values indicate threshold estimation errors using parametric bootstrapping procedures.
Figure 2.
Comparisons of REPT and MOBS performance in terms of accuracy (Panel A) and variability (Panel B) for 10 participants.
Panel A shows the difference between REPT and MOCS on the horizontal axis and the difference between MOBS and MOCS on the vertical axis, for each participant. It is clear that the REPT-MOCS values are closer to zero than are the MOBS-MOCS values, as confirmed by the fact that the absolute difference between REPT and MOCS was significantly smaller than the absolute difference between MOBS and MOCS. This suggests that REPT thresholds were more accurate than MOBS thresholds. Panel B displays the relationship between the standard deviation of REPT and MOBS thresholds for each participant. The variability of phosphene thresholds measured with REPT was significantly lower than the variability of phosphene thresholds measured with MOBS which suggess that REPT thresholds were more consistent.
Figure 3.
Scatterplot between REPT and MOCS (Panel A), and MOBS and MOCS (Panel B) phosphene thresholds for 10 participants.
Both REPT and MOBS thresholds well correlated with the phosphene thresholds estimated using MOCS. However, the relationship between REPT and MOCS appears stronger than between MOBS and MOCS.
Figure 4.
The duration of REPT and MOBS procedures.
Panel B shows the average duration of the REPT and MOBS procedures for each participant. The duration of the REPT was significantly shorter compared with the duration of the MOBS. The duration of REPT was also more predictable, compared with MOBS, because it always finished in 30 trials, while the number of trials in MOBS varied considerably. Error bars represent 95% bootstrapped confidence intervals.
Figure 5.
The effect of stimulation frequency on phosphene thresholds using REPT for 7 participants.
This scatterplot shows mean phosphene thresholds for each participant estimated using REPT with the average stimulation frequency of 0.36 Hz on the horizontal axis and 0.16 Hz on the vertical axis. The diagonal line indicates where the two estimates are equivalent. No statistically significant difference between the slow and fast staircases was found.