Bak et al. (1990) recorded threshold minima at depths of 2–3 mm, 4 mm and 4.5 mm in three sighted volunteers undergoing occipital craniotomy for excision of epileptic foci. In the patient with the lowest detection thresholds, they plotted the threshold stimulus current vs. electrode depth, showing the lowest thresholds (20 µA) at a depth of approximately 2.25 mm.

In their subsequent study on a blind volunteer, the same group reported thresholds varying from 1.9 µA to 77 µA using fixed-length electrodes implanted to a depth of 2 mm (Schmidt et al., 1996). As noted by Torab et al. (2011), the undulating nature of the cerebral cortex renders it difficult to ensure consistent penetration depth of all electrodes with an array based on a rigid substrate. Y-27632 concentration Moreover, the ability of electrodes to elicit behavioral responses at current levels not damaging to the electrodes or tissue may be predicated partly on the location of electrode stimulating sites within

laminae containing the most excitable neuronal elements. Spatial differences in threshold current (DeYoe et al., 2005) or depth of lowest threshold (Bak et al., 1990) and natural variations in the thickness of V1 (Fischl and Dale, 2000) may therefore combine to present a significant challenge for ensuring implantation of electrodes to the optimal depth in visual cortex. Possible solutions to these problems include the implantation of arrays with electrode shanks of Caspase phosphorylation varying length as previously

described (Bradley et al., 2005), which may require an increase in the density of electrodes, e.g. (Wark et Cyclooxygenase (COX) al., 2013) to preserve the resolution of the phosphene map. Another possible solution could be the incorporation of multiple stimulating sites onto individual electrode shanks (Changhyun and Wise, 1996) or microdrives that allow independent adjustment of electrode penetration depth (Gray et al., 2007, Yamamoto and Wilson, 2008 and Yang et al., 2010). For the latter, further reductions in the size of the positioning hardware will be required before integration into high electrode count arrays is a realistic possibility. Reductions in the size of electrode arrays may also offer some benefits; for example, the Illinois group and EIC Laboratories recently described a 2×2 mm, 16-electrode array (Kane et al., 2013) that may permit improved consistency of electrode tip depth relative to the curved cortical surface when implanted over a wide area. One potential disadvantage to this approach is the larger number of arrays to be implanted, and its potential implications for the length of the surgical procedure. For example, implanting 650 electrodes in groups of 16 would require approximately 41 arrays (Srivastava et al., 2007), while implanting 500 electrodes in groups of 43 would require only 11 (Lowery, 2013).