BT2 Q&A
Q: Why are the settings (filters) for different frequency channels different? Butterworth/Elliptic
A: Elliptic filters typically have a sharper roll-off, which allows for more precise measurement of a specific frequency range. This is especially useful in some situations where we want the rewards (Rew) to stay within certain values, as some people may be quite sensitive even to small variations in the reward frequency, depending on the protocol.
For inhibits, we prefer using Butterworth filters because they have a faster response, albeit with a more gradual roll-off. Since inhibits usually involve broader frequency bands and do not require as much precision, it’s acceptable if the filter captures a bit of activity slightly outside the set limits.
Q: I was not able to understand the shadow panel limits (calibration) – even when all targets were met, the screen was darkened. But then also, sometimes I forced all targets not to be met, and the screen hardly darkened.
A: The shadow feedback has different operating modes. Some are continuous modes, while others are contingent (yes/no type). The contingent feedback options are “Number of Met Targets” and “Binary.” In these cases, the behavior should match what you expected: if all targets are met, the screen becomes bright; if not, it darkens.
The key difference between “Number of Met Targets” and “Binary” is that the former adjusts the shadow’s brightness in gradations, depending on how many targets are being met, whereas the latter simply switches between a fully bright screen (when all targets are met) and a fully dark screen (when at least one target is not met).
In continuous modes, the behavior is quite different. Instead of checking whether each target is met, they look at how the current value of each metric compares to the threshold. For example, if an inhibit has a threshold at value 5 (arbitrary units), and the current value is 6, the screen may darken slightly. If the value increases to 8, the screen will become even darker. If it decreases to 4, the screen will brighten — and even more so if it reaches 3.
Thus, the shadow’s brightness varies continuously in response to the moment-to-moment values of all training parameters, and it may never be fully bright nor at maximum darkness.
Additionally, in some composite protocols (such as FRE Peak Frequency, BAL Multiplex, and CON Coherence), the shadow is programmed to respond only to certain parameters, while other parameters may influence only other feedback elements like sounds.