The actual binaural receptive fields ( Figures 4C–4E) do not have this simple form, revealing the complex, frequency-dependent, binaural tuning of the subthreshold inputs. The ability to measure the inputs to the MSO neurons in vivo allowed us to test how inputs from both ears sum. To this end, we compared the measured averaged response during the beat cycle with the prediction from a purely linear interaction of the monaural contributions obtained by averaging across the respective ipsi- and contralateral tones (Figure 5A; Movie S1). The observed responses closely followed the linear prediction, which accounted

for 97.9% of the variance. The success of the linear prediction was a general finding, and was observed for both juxtacellular and whole-cell recordings (Figures 5B and S7). Careful inspection of the LDK378 datasheet raw traces did not reveal fast, downward going events that specifically preceded the positive events, both in whole-cell and in juxtacellular recordings (Figures 2B and S3). Simultaneous juxta- and whole-cell slice recordings indicated that the resolution of the juxtacellular recordings allows detecting IPSPs with an amplitude < 1 mV (Figure S2). We therefore did not find evidence for well-timed inhibition, nor for a substantial effect of the current sink presented by the nonstimulated dendrite. To further test this linearity, we compared the binaural beat response with the responses to monaural

stimulation using the same tones as in the binaural beat stimuli (Figure 5C). Summing the monaural responses provided an excellent prediction of the binaural responses (Figure 5D), accounting for Galunisertib clinical trial 95.5% of the variance.

Rolziracetam The small deviations are analyzed in Figure S7. Previous extracellular recordings from MSO have shown that firing rate at the “worst ITD” is generally lower than the rates obtained by monaural stimulation of either ear, and can even drop below the spontaneous rate (Goldberg and Brown, 1969; Spitzer and Semple, 1995; Yin and Chan, 1990). We observed that subthreshold responses were highly stereotyped, repeating themselves each beat cycle. We therefore determined not only the mean subthreshold potential (Figure 5A), but also the variance across beat cycles (Figure 5E). The across-cycle variance varied systematically during the cycle. It was clearly larger when responses were large, but in between, during 64% of the beat cycle, it systematically dropped below the spontaneous level. At its absolute minimum, it amounted to only 5% of the spontaneous level. Especially since the inhibitory inputs are large and few (Couchman et al., 2010), the deep trough of the across-beat-cycle variance appears to signify an absence of excitatory inputs rather than the presence of well-timed inhibition. More examples are shown in Figure 6. The periodic reduction of the variance below the spontaneous value was observed in all 22 ITD-sensitive cells.