NeuroAnalyzer tutorials: Statistics: Bootstrapping

using NeuroAnalyzer
using Plots
using Statistics
eeg = load("files/eeg.hdf")
e10 = epoch(eeg, ep_len=10)
delete_epoch!(e10, ep=1:10);

Calculate and plot CI using bootstrapping

Get the first 2 seconds of the channel 1 of epochs 1:10 and calculate its 95%CI:

s = e10.data[1, 1:512, 1:10]
t = e10.epoch_time[1:512]
s_avg, s_l, s_u = bootstrap_ci(s, cl=0.95)

(s_avg = [11.659300281192465, 10.267416842024419, 11.982210796098316, 12.633272561174595, 11.574590513319876, 11.545503015764897, 10.547236039222913, 10.538442595517484, 10.428070967561805, 10.747234389826788  …  -14.947999572996457, -17.49806253030125, -16.394520781528968, -12.147031307974233, -9.684243826578012, -12.314498061380473, -15.23403828747199, -15.212713925487574, -12.856349899754504, -11.02106059836681],
 s_ci_l = [10.832015791592202, 9.425955184854137, 11.113173319624332, 11.655034156537877, 10.670701358751547, 10.662448915206648, 9.678479041456328, 9.769467078606432, 9.567759547887205, 9.855897809182952  …  -17.191546471900878, -20.275959660861844, -19.094147147668767, -14.231077682935002, -11.498618960564453, -14.548026289259298, -17.991393034019794, -17.924602552610367, -15.010052229831853, -12.922218878394549],
 s_ci_h = [12.507459320826188, 11.134337864508433, 12.881451210821625, 13.63029221910113, 12.500960097313936, 12.473436756802954, 11.437148045689197, 11.353501323071535, 11.343948770430805, 11.694571928671545  …  -12.856892172501105, -14.892851740964339, -13.824024518626823, -10.13576466727993, -7.915514309518437, -10.189946272576654, -12.612162419520475, -12.650293236138118, -10.792592966626925, -9.187550954945998],)

Plot CI:

plot_ci(s_avg, s_l, s_u, t)

Calculate statistic using bootstrapping

Any statistical function may be calculated using bootstrapping method. The function must be provided as f="function_name(obj)", obj will be replaced with the signal data, e.g. f="mean(obj)".

s = e10.data[1, :, 1:10]
s_stat = mean(s)
s_dist  = bootstrap_stat(s, f="mean(obj)")

3000-element Vector{Float64}:
  0.005161853605707733
 -0.03585747533203385
  0.15995404814361916
  0.08269136350010058
  0.11693123097306657
 -0.057481567153723746
  0.09996914964729013
  0.2512558627746825
  0.21215424722436368
  0.07628844706772409
  0.1921110686558194
  0.07281927963930865
  0.07271812863405688
  ⋮
  0.16344561232525195
  0.06386297452225129
  0.15517305104924634
 -0.009586425558272538
 -0.013604292655014127
  0.03397578181076675
  0.22080657633693876
  0.13258182707720811
  0.2266256048958903
 -0.01567515394520047
  0.009050636357872044
  0.13110245120866698

plot_histogram(s_dist,
               s_stat,
               draw_median=false,
               draw_mean=true)