The following image shows the transient response of the Atlantis/Discovery amplifiers to a single, sudden "shift" in the DC level of the input. The input was suddenly shifted by 20 microvolts, then kept at the new level. It did NOT oscillate, and the system did NOT take "minutes" to respond. It responded in seconds. The response of the amplifier, as well as of each frequency band component, is clearly visible. This is NOT an oscillation or a frequency or an AC signal. It is a DC shift. Because it is a "step" function, it contains a multitude of frequencies, all of which show up. We are seeing the "step" response of the filters. Those who think in terms of "oscillations" and Fourier Transforms may miss the point of the step response, which requires a Laplace Transform to properly measure. However, even the bandpass filters do show a "bump" or "ringing" that can be detected. This "ringing" step response is well-known to electrical engineers. By setting the filters at very low cutoff frequencies, they become oblivious to all the midband frequencies, and only show the "shift", however attenuated or delayed they may be. We also show an "Othmer" filter set with a bottom of 0.01 which is close to the single-pole rolloff of the NeuroAmp. For comparison, we show a signal with a bottom of 0.001 to show how it responds similarly, but with a larger signal. This shows that a transient shift in the DC level of the signal can be detected even with AC bandwidths, but that the Atlantis, which is a DC amplifier, has a larger as well as a more rapid, response.
For those who want to know how neurons can produce these types of DC shifts, I have attached a few pages from Neidermeyer and Lopes da Silva (2005) showing that pyramidal cells and EPSP's can indeed produce shifts lasting minutes or hours. One figure is appended at the end of this article as well.