As anticipated, all channels where neurons were located over electrodes responded to the bath ground. We considered a regional application relying on interconnectivity for signal radiation. To accomplish this regional signaling, we utilized a series of 3 adjacent electrodes for the entire circuit, including one of the array electrodes as the promixal ground, rather than relying on the bath ground (Fig. 1B). In the example presented in Fig. 1B, we stimulated electrode 12, recorded from adjacent electrode 11 and utilized a local electrode 10 as the ground electrode. This setup allowed regional stimulation and analyses, and could be situated anywhere within the array. In preliminary studies, we observed that disconnecting the bath ground resulted in excessive noise that precluded reproducible signal monitoring. Therefore, we retained the bath ground, but minimized the extent of current sinking at the bath ground by selecting an array electrode for hard-grounding that was situated between the stimulus and bath ground electrodes (Fig. 1B). Hard-grounding was required to minimize the circuital impedance of the reference electrode and insure that it would not exceed that of the bath ground. The appropriate stimulus access pin on the amplifier (in this case the pin for electrode 10; Fig. 1B) was connected to signal ground using a wire lead comprised of a mini-test clip and an alligator clip.
We then applied the square wave and Biological Signal using this local arrangement. (Fig. 1C). Both the square wave and the Biological Signal invoked similar overall numbers of response signals (37 ± 3 vs 34 ± 1, respectively in a 30 sec interval, n = 10 each), which, rather than radiating throughout the culture, were confined to those electrodes connected via synaptic connections to the localized stimulation and response electrodes (not shown; see reference 13). Notably, the Biological Signal induced relatively more bursts (defined as continuous responses that remained above baseline for ≥ 3sec; 28-30) than individual signals (p<0.03; Student's t test; Fig. 1C).
Since prior studies indicated that improved response was observed by sequential stimulation at 2 adjacent electrodes (31), we also applied the Biological Signal at 2 sets of adjacent electrodes (electrodes 15 and 12; and electrodes 12 and 7), and recorded responses from a third electrode situated midway between these pairs (electrode 11). Since the Biological Signal was not a simple biphasic signal, this also afforded us the opportunity to apply it in an inverted form. Application of the Biological Signal at electrodes 15 and 12 did not invoke a response, but did evoke a response when applied at electrodes 12-7. By contrast, the inverted Biological Signal invoked a stimulus when applied at electrodes 15-2 but not at electrodes 12-7 (Fig. 1D). These differential responses from the same electrodes rule out any underlying electrical artifact, since inversion of the signal determined whether or not a response was observed.
In those instances where the Biological Signal invoked a response, it occurred within a similar interval as observed with either single or triple square wave pulses. Spontaneous signaling resumed within a statistically-identical interval after stimulation with the Biological Signal when compared to square wave pulses (4.2 ± 0.6 sec). Notably, in those instances where the Biological Signal did not invoke a response, we observed a doubling of the delay in resumption of spontaneous signals; spontaneous signals resumed after 8.1 ± 1.2sec. This is approximately twice the delay observed following all other stimulation conditions and approximately twice the normal interval between spontaneous signals (above). One interpretation of this phenomenon is that the Biological Signal generates an inhibitory signal under some conditions. These results further indicate that, depending upon signal orientation and location, the Biological Signal can invoke inhibitory or excitatory responses that alter downstream neuronal signaling.
The localized stimulation system described herein requires minimal modification of exisiting approaches and no harware alterations. It provides localized stimulation with a specificity approaching that of microarrays using predefined tunnels that restrict the directionality of axonal outgrowth (and therefore restrict the directionality of signal propagation (32). Our altered signaling and circuit design is readily reversible to classical settings, which can allow localized versus global stimulation and monitoring of the same neuronal networks. Stimulation with a recorded synaptic signal may be more effective than classical square waves in attempting to strengthen connections within cultures or evoke activity in otherwise inactive subpopulations. Notably, the efficacy of a pre-recorded synaptic signal as described herein provides the novel opportunity for comparision of the effects of stimulation with normal versus aberrant synaptic signals (e.g., epileptiform signals or those elicited in response to excitoxicity, etc.). An additional modification could include the utilization of an external reference electrode. These possibilities will be examined in future studies.