1) allanimaldata.mat - processed data

• version – version number for this format of data storage
• freqrange – bandpass filter range (Hz) for the EMG
• impedances – impedance values (Ω) for electrode channel
• animalID – unique Sober lab ID tag for animal
• muscles – muscle location for each electrode channel
• implantDate – date (yyyy-mm-dd) of electrode implant surgery
• experimentDate – date (yyyy-mm-dd) of treadmill running experiment
• experimental_rig – treadmill name
• ntrials – number of trials
• exclusion_trials – trials pre-excluded manually due to issues
• comments – notes on experiment
• units – 3xm matrix with info on m units (muscle location for each unit, noise floor on the most relevant channel for each unit, signal amplitude for each unit on that same channel)

trialdata – each trial consisted of 1 minute of continuous running on the treadmill at a given speed

• trialnum – trial number
• beltSpeedL – left belt speed (m/s) for whole trial duration
• beltSpeedR – right belt speed (m/s) for whole trial duration
• right_forepaw_strides – stride times for the right forepaw (Each paw has an nx4 matrix ordered footstrike₁, liftoff, footstrike₂, stride number (n))
• elb_angle_min_stride – re-defined strides based on the elbow minimum angles because it better captures spike patterns from the elbow extensor muscles. This was only calculated for the right elbow. The nx5 matrix has n strides found through footstrike/liftoff definition, and columns are ordered as timestamps (s) for elbow minimum₁, footstrike₁, liftoff, elbow maximum, elbow minimum₂. Since the elbow angle cannot always be reliably calculated, some strides have already been excluded
• spikes_times – spike times for each unit. Units have been ordered previously from most to least total spikes

stride_discrete_kinematics – discrete kinematic parameters for 6 experimental mice. Each field contains n rows based on the right forepaw defined strides of the animal

• stride_dur – stride duration (s) as time between footstrike₁ and footstrike₂
• stance_dur – stance duration (s) as time between footstrike₁ and liftoff
• swing_dur – swing duration (s) as time between liftoff and footstrike₂
• elb_stride_dur – stride duration (s) as time between elbow minimum₁ and elbow minimum₂
• elb_stance – duration (s) between elbow minimum₁ and footstrike₁
• vel – body velocity (m/s) averaged over the duration from footstrike to footstrike
• accel – body acceleration (m/s²) averaged over the duration from footstrike to footstrike
• orientation – distance (m)between nose and tail, averaged over the duration from footstrike to footstrike – used to validate orientation, as values should be positive if the animal is walking forwards
• speed_quart – nx4 matrix with n strides and true values in a single column based on evenly dividing all valid strides into 4 groups based on speed. 1^st column is slowest and 4^th is fastest. This definition (which is the main one used) can be helpful to account for the natural variability of animals, with analyses based on equal group sizes
• speed_range – nx4 matrix with n strides and true values in a single column based on pre-determined group speeds. 1^st column is 0.1-0.15 m/s. 2^nd is 0.15-0.2 m/s. 3^rd is 0.2-0.25 m/s. 4^th is 0.25-0.3 m/s. This definition can be helpful for consistency across animals
• speed_tread – first column is treadmill speed during each stride. The second column is whether body velocity falls within 90-110% of the treadmill speed.

stride_discrete_spiking – discrete spiking parameters for 6 experimental mice. Each field is a nxm matrix with n rows based on right forepaw strides of the animal and m columns based on number of units

• spike_count – number of spikes in each unit per stride
• active_duration – duration (s) between first and last spikes within a stride
• active_duration_zeros – duration (s) between first and last spikes within a stride where each spike is artificially given a duration of 4ms around the peak. Thus, strides with one spike have a time of 4ms instead of being excluded
• duty_cycle – active duration divided by elbow stride duration,
• ifr_med – median value (Hz) of the instantaneous firing rate for all spikes. Instantaneous firing rate was calculated as 1 divided by the inter-spike interval
• ifr_max – maximum value (Hz) of the instantaneous firing rate for all spikes. Instantaneous firing rate was calculated as 1 divided by the inter-spike interval
• first_spk – time (s) of first spike
• last_spk – time (s) of last spike
• fr_maxraw – maximum firing rate (Hz) within stride from firing rated defined by gaussian smoothed over spike times
• fr_phaseraw – approximate phase of maximum firing rate based on scaling continuous firing rate trace over time to a 1-100 phase scale. Note that values of 1 are almost always erroneous from strides with no spiking from a unit

2) allmousekinematics.mat

Timeseries data for kinematic tracking of each mouse during the experiment. Each mouse's data is located in order in a table with the following information

• beltspeed - vector of treadmill belt speed across all trials
• session_time - vector of timepoints across all trials
• trial_number - trial number
• kinematics - labeled body parts with 'prob' representing the confidence of the tracked point along (x,y) dimensions. If choosing to look at the video files directly, skipped frames sometimes occur and can be identified in each trial based on when a confidence point has a NaN value.

3) mouseRawData.mat

Contains processed EMG data and timeseries information for mice A-F in struct trialdata_original with the following fields:

• emg - mxn EMG data matrix of m data samples recorded at 30kHZ and n electrode channels
• time - mx1 time vector
• trial - start and stop indices for each trial

4) All_videos.zip

Locomotion videos for each mouse (mouse A to F) and trial contained within this dataset and the video files are named based on number of trials.