# Cerebellum Dataset

% Data Repository related to the paper "Cerebellar activity predicts vocalization in fruit bats"
% (currently under minor revision in Current Biology).
%
% The data is stored in .mat files containing neural recordings from the cerebellum of fruit-eating bats,
% captured during the production of both echolocation and social calls.

## Data Guide: Cerebellar Activity Predicts Vocalization in Fruit Bats
%
% This repository contains data collected from cerebellar recordings in fruit-eating bats during the
% production of echolocation and social calls. The data is structured into various .mat files, with details of
% each dataset and its specific parameters explained below.

%% 1. *Frequency Tuning LFP
% This dataset includes Local Field Potentials (LFP) data captured from 224 channels, organized by stimulus
% level and frequency.

*FT_LFP:
    * ft: Spike count (224 channels x levels x frequency)
    * besterp: Evoked response data across levels (30:15:90) and frequencies (15:5:85)
    * bestresponse: Best frequency-band level response (BFBL) for each channel
    * cfmt: Characteristic Frequency/Minimum Threshold for each channel (CFMT)

%% 2. *Frequency Tuning Spikes
% This dataset contains spike count data from 224 channels, organized similarly by stimulus level and frequency.

*FT_Spikes:
    * ft: Spike count (224 channels x levels x frequency)
    * N2w: Peri-Stimulus Time Histogram (PSTH) data across levels (30:15:90) and frequencies (15:5:85)
    * bestresponse: Best frequency-Best level response (BFBL) for each channel
    * cfmt: Characteristic Frequency/Minimum Threshold for each channel (CFMT)

%% 3. *Vocs (Vocalization Data)
% This dataset provides information about the call parameters and classification data for both echo and communication calls.

*callparams:
    * dur2: Duration of calls, categorized by call type
    * pf2: Peak frequency of calls, categorized by call type
    * finalMatrix: Cross-correlation matrix of different call types
    * newClass: Identifiers for calls used in cross-correlation

  

*SpikeClassifData:
    * Ne: Normalized echo call spike data for training the SVM
    * Nc: Normalized communication call spike data for training the SVM
    * N2: Spike count data
    * N3: Identifiers for spike count data


 *LFPClassifData:
    * echoData: LFP data during echo calls for training the SVM
    * commData: LFP data during communication calls for training the SVM

% ---
% This guide outlines the structure of the data, which is organized into several .mat files.
% MATLAB version: R2021a.