Importantly, the frequency content in the output of a demodulating system will not depend on the carrier TF. Responses to interference patterns could also result from nonlinear (multiplicative) interactions between the different component frequencies present in the stimulus. The possible nonlinear interactions are limited by the observation that Y cell responses to interference patterns with

a static carrier contain power at the envelope TF and twice the envelope TF (Demb et al., 2001b and Rosenberg et al., 2010). The simplest nonlinear interaction that would explain this observation is the sum of pairwise multiplications selleck inhibitor of the component frequencies. In response to a three component interference pattern, this nonlinearity would produce five dominant response frequencies: (1) TFenv, (2) 2TFenv, (3) 2TFcarr, (4) 2TFcarr – TFenv, and (5) 2TFcarr + TFenv. Note that with a static carrier, the only response components are at TFenv and 2TFenv, as previously observed experimentally. Nonlinear interactions such as these may result in responses at the envelope TF,

but the responses are not demodulated since they also include a set of carrier-dependent output frequencies. For instance, carrier-dependent responses are observed in the output of individual hair cells in the peripheral auditory system (Jaramillo et al., 1993). Because the carrier was Sirolimus in vitro held static in previous Y cell experiments, demodulating and nondemodulating nonlinearities could not be differentiated. Importantly, the frequency content in the

output of a non-demodulating nonlinear whatever system will depend substantially on the carrier TF. It is thus possible to differentiate a demodulating system from a linear or other nonlinear system by presenting interference patterns at different carrier TFs and examining the frequency content in the output. To determine the frequency content in Y cell responses to interference patterns, peristimulus time histograms (PSTHs) with 10 ms bins were constructed and mean subtracted. Power spectra were then computed from the fast Fourier transforms of the PSTHs and each power spectrum was normalized to have a maximum value of one. For each carrier TF, a population averaged power spectrum was then calculated using responses to interference patterns with the same envelope TF (5.6 cyc/s). Regardless of the carrier TF, the responses oscillated predominantly at the envelope TF (Figure 3). Progressively smaller but distinct peaks attributable to static (e.g., half-wave rectification and expansive) nonlinearities inherent to spiking neural responses were also observed at the second and third harmonics of the envelope TF. Similar response patterns were observed at both lower and higher envelope TFs (Figure S1). Thus, the frequency content in Y cell responses to interference patterns does not depend substantially on the carrier TF.