Bowman et al MS data.xlsx

Time to loss of visual indicators and EEG signal data from 9 rainbow trout. EEG signals were recorded using a custom designed, non-invasive electrode fitted silicone cup attached to the heads of rainbow trout using a peristaltic pump.<br>EEG signals were recorded for 10 min with light stimulus. Light stimulus was delivered at 2 Hz using an LED strobe-light in a dark room. The light was detected using an in-house designed custom made light detector connected to a PowerLab system. After 10 min of stimulus, the peristaltic pump was turned on high for 30 sec to ensure the cup was still securely in place and reduce the chance of the cup coming off in the treatment bath. The fish was then gently manually transferred to a 70L tank containing 150 mg L^-1 MS-222 and 300 mg L^-1 sodium bicarbonate in 51 L of 10°C water. EEG was recorded during exposure to anesthetics for 30 min with light stimulation. EEG recording was paused in between each time period (10 min with light, 30 min treatment with light) to eliminate artifacts in the signal caused by cup adjustment, turning equipment on or off, or transferring the fish between tanks. Room lights were off for the entirety of the experiment. <br>EEG signals were analyzed using LabChart 7 Pro (version 7.3.8, AD Instruments, Cologne, Germany). A 0.5-32 Hz bandpass filter was used to help eliminate electrical noise, noise from movement, and signal irregularities. 120 consecutive, non-overlapping, half-second time windows were averaged to give a single averaged 500 ms signal representative of each minute of recording. Averaged signals consisting of less than 60 half-second windows (i.e. &lt;50% of successful recordings during 1 min of recording) were omitted from analysis as they may not provide an accurate measurement. For analysis, EEG recordings were filtered into high frequency, low amplitude (8-32 Hz) beta and alpha waves and the low frequency, high amplitude (0.5-8 Hz] theta and delta waves for analysis. VERs were visually determined as present or absent in the beta wave (13-32 Hz) where they were most distinct. VERs were considered lost when the VER wave amplitude was indistinguishable from the amplitude of the rest of the wave (Figure 4). <br> The EEG signal (0.5-32 Hz) amplitude was measured in LabChart as the maximum point of the wave minus the minimum point for each representative minute of recording. Time until signal amplitude reduced to &lt;50% and &lt;12% pre-treatment amplitude was calculated for each fish using Microsoft Excel (2016). To assess shifts between waves, median frequency and percentage of total power (Ptot) were calculated from the EEG signal (0.5-32 Hz) in LabChart, using a fast Fourier transform (FFT) with a Hann (cosine-bell) window with 50% overlap in LabChart. The EEG signal was separated into high frequency, low amplitude (8-32 Hz) beta and alpha waves and low frequency, high amplitude (0.5-8 Hz] theta and delta waves for Ptot analysis. EEG signal amplitude, median frequency, and Ptot were analyzed using a linear mixed model with AR(1) as the covariance structure and time as the repeated variable in SPSS. The mixed model was used to determine if there was a significant difference between the averaged measurements taken during the pre-treatment period compared to each representative minute of treatment. When multiple tests were performed p-values were adjusted using the Bonferroni method. Amplitude data was log transformed to meet the assumptions of the model.