Gravity EMG Sensor sampling question

I have the Gravity EMG sensor working with the provided filters (and some i've written) to get some basic controls working as a proof of concept prosthetic control system in the video below. Now i'd like to broaden the use case by using some machine learning models to detect more muscle signals than simple conditional programming to detect muscle signals within thresholds. I have a model trained to detect weak, medium and strong flexes with 92% accuracy in the Edge Impulse platform to deploy onto an Arduino Nano 33 BLE Sense. I'm working on deploying the trained model back to the Arduino, but need to wrap my head around the how the sensor sampling rate works to fill the buffer on which the classification is based. I am using your EMGFilter library to get the filtered signal (which as I understand filters out the power line frequency - 50 or 60 Hz - and samples at either 500 or 1000 Hz). I am using 60 Hz (i'm in the US) and 1000 Hz (for more data) and have read this as the input into Edge Impulse for training/testing data. I think my confusion is coming from Edge Impulse but want to be sure I understand the sensor before reaching out to them. Their system automatically detects the sampling frequency of any connected sensor and it detects a sample rate of 4000 Hz. I am SURE that my sampling rate is set to 1000 Hz as I'm successfully using your filters to get my filtered signal into Edge. I can't see anything in the EMGFilters library that suggest the data is resampled to a 4000 Hz frequency…any ideas?