wonambi-python.github.io/gui/analysis.html at master · wonambi-python/wonambi-python.github.io

History

453 lines (438 loc) · 44 KB

Raw

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

<!DOCTYPE html>

<head>

<title>Analysis console — wonambi 7.11 documentation</title>

<!--[if lt IE 9]>

<![endif]-->

</head>

<a href="../index.html" class="icon icon-home"> wonambi

</a>

7.11

</div>

</form>

</div>

</div><div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="Navigation menu">

<li class="toctree-l1"><a class="reference internal" href="../introduction.html">WONAMBI</a><ul>

<li class="toctree-l2"><a class="reference internal" href="../introduction.html#features">Features</a></li>

<li class="toctree-l2"><a class="reference internal" href="../introduction.html#installation">Installation</a></li>

<li class="toctree-l2"><a class="reference internal" href="../introduction.html#documentation">Documentation</a></li>

<li class="toctree-l2"><a class="reference internal" href="../introduction.html#change-log">Change Log</a></li>

<li class="toctree-l2"><a class="reference internal" href="../introduction.html#requirements">Requirements</a></li>

<li class="toctree-l2"><a class="reference internal" href="../introduction.html#optional-requirements">Optional Requirements</a></li>

<li class="toctree-l2"><a class="reference internal" href="../introduction.html#status">Status</a></li>

<li class="toctree-l2"><a class="reference internal" href="../introduction.html#license">License</a></li>

</ul>

</li>

<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a><ul>

<li class="toctree-l2"><a class="reference internal" href="../installation.html#linux-mac-os-x">Linux / Mac OS X</a></li>

<li class="toctree-l2"><a class="reference internal" href="../installation.html#windows">Windows</a></li>

</ul>

</li>

<li class="toctree-l1"><a class="reference internal" href="../analysis/index.html">Tutorial / Analysis</a><ul>

<li class="toctree-l2"><a class="reference internal" href="../analysis/tutorial.html">Get Started</a></li>

<li class="toctree-l2"><a class="reference internal" href="../analysis/spectrum.html">Frequency Domain</a></li>

<li class="toctree-l2"><a class="reference internal" href="../analysis/plot3d.html">Plot 3D images</a></li>

</ul>

</li>

<li class="toctree-l1"><a class="reference internal" href="../cmd.html">Command Line</a><ul>

<li class="toctree-l2"><a class="reference internal" href="../cmd.html#wonambi">wonambi</a></li>

<li class="toctree-l2"><a class="reference internal" href="../cmd.html#won-convert">won_convert</a></li>

</ul>

</li>

<li class="toctree-l1 current"><a class="reference internal" href="index.html">Graphical User Interface</a><ul class="current">

<li class="toctree-l2"><a class="reference internal" href="open.html">Open Dataset</a></li>

<li class="toctree-l2"><a class="reference internal" href="labels.html">Edit Labels</a></li>

<li class="toctree-l2"><a class="reference internal" href="channels.html">Plot Channels</a></li>

<li class="toctree-l2"><a class="reference internal" href="traces.html">Navigation</a></li>

<li class="toctree-l2"><a class="reference internal" href="notes.html">Annotations</a></li>

<li class="toctree-l2"><a class="reference internal" href="detect.html">Event detection</a></li>

<li class="toctree-l2"><a class="reference internal" href="methods.html">Detection methods</a></li>

<li class="toctree-l2 current"><a class="current reference internal" href="#">Analysis console</a></li>

<li class="toctree-l2"><a class="reference internal" href="screenshot.html">Screenshot</a></li>

</ul>

</li>

<li class="toctree-l1"><a class="reference internal" href="../api/wonambi.html">Modules, classes, functions</a><ul>

<li class="toctree-l2"><a class="reference internal" href="../api/wonambi.html#subpackages">Subpackages</a></li>

<li class="toctree-l2"><a class="reference internal" href="../api/wonambi.html#submodules">Submodules</a></li>

</ul>

</li>

<li class="toctree-l1"><a class="reference internal" href="../testing.html">Development</a><ul>

<li class="toctree-l2"><a class="reference internal" href="../testing.html#prepare-test-environment">Prepare Test Environment</a></li>

<li class="toctree-l2"><a class="reference internal" href="../testing.html#get-files">1. Get Files</a></li>

<li class="toctree-l2"><a class="reference internal" href="../testing.html#run-the-tests">2. Run the Tests</a></li>

<li class="toctree-l2"><a class="reference internal" href="../testing.html#coverage">3. Coverage</a></li>

<li class="toctree-l2"><a class="reference internal" href="../testing.html#documentation">4. Documentation</a></li>

<li class="toctree-l2"><a class="reference internal" href="../testing.html#release">5. Release</a></li>

<li class="toctree-l2"><a class="reference internal" href="../testing.html#tips-and-tricks">Tips and Tricks</a></li>

</ul>

</li>

<li class="toctree-l1"><a class="reference internal" href="../changelog.html">Change Log</a><ul>

<li class="toctree-l2"><a class="reference internal" href="../changelog.html#version-7">Version 7</a></li>

<li class="toctree-l2"><a class="reference internal" href="../changelog.html#version-6">Version 6</a></li>

<li class="toctree-l2"><a class="reference internal" href="../changelog.html#version-5">Version 5</a></li>

<li class="toctree-l2"><a class="reference internal" href="../changelog.html#version-4">Version 4</a></li>

<li class="toctree-l2"><a class="reference internal" href="../changelog.html#version-3">Version 3</a></li>

<li class="toctree-l2"><a class="reference internal" href="../changelog.html#version-2">Version 2</a></li>

<li class="toctree-l2"><a class="reference internal" href="../changelog.html#version-1">Version 1</a></li>

</ul>

</li>

</ul>

</div>

</nav>

<a href="../index.html">wonambi</a>

</nav>

<li class="breadcrumb-item"><a href="index.html">Graphical User Interface</a></li>

<li class="breadcrumb-item active">Analysis console</li>

</li>

</ul>

<hr/>

</div>

<h1>Analysis console<a class="headerlink" href="#analysis-console" title="Permalink to this heading"></a></h1>

Wonambi’s analysis console allows the flexible selection of signal for a variety of analyses, including frequency domain analyses and phase-amplitude coupling (PAC).

The console is read from left to right.

On the left-hand half of the console is the signal selection pane.

Signal can be selected by event, epoch or longest run, and by channel, cycle and stage, with flexible concatenation options, and with artefacted signal exclusion.

On the right-hand half is the analysis pane, arranged into 3 tabs: Frequency, PAC and Events.

You may apply any or all of these analyses at once by activating them in each tab.

To open the dialog, click on <code class="docutils literal notranslate">Analysis</code> -> <code class="docutils literal notranslate">Analysis console</code>.

<h2>File location<a class="headerlink" href="#file-location" title="Permalink to this heading"></a></h2>

Select the base name and location of the data files.

The analysis console creates CSV files containing the raw analysis data, with suffixes for each analysis type:

<li>_freq: Frequency</li>

<li>_evtdat: Events</li>

<li>_fooof: Spectral parametrization report</li>

</ul>

Note

These data files can become quite large depending on the analysis.

</div>

</section>

<h2>Chunking<a class="headerlink" href="#chunking" title="Permalink to this heading"></a></h2>

Different analyses require different lengths of signal, hence the chunking option. You may chunk by <code class="docutils literal notranslate">event</code>, <code class="docutils literal notranslate">epoch</code> or <code class="docutils literal notranslate">longest run</code>.

Chunking <code class="docutils literal notranslate">by event</code> gathers all events of the type(s) selected as individual segments.

If you wish to gather only the signal for the channel on which an event was marked, keep the <code class="docutils literal notranslate">Channel-specific</code> box checked.

Alternatively, you may wish to gather signal concurrent to an event on all selected channels.

In this case uncheck <code class="docutils literal notranslate">Channel-specific</code>.

Note

<code class="docutils literal notranslate">Channel-specific</code> will only gather events marked on the specified channels.

If unchecked, all events of the specified types will be gathered, including those

marked on channels other than those specified, and including “orphan” events

with no associated channel.

</div>

Chunking <code class="docutils literal notranslate">by epoch</code> gathers all relevant signal cut up into segments of equal duration.

To obtain the epochs as segmented by the Annotation File for sleep scoring, check <code class="docutils literal notranslate">Lock to staging epochs</code>.

Alternatively, you may select a different epoch length with <code class="docutils literal notranslate">Duration (sec)</code>.

In this case, all relevant signal, after the specified concatenation, will be segmented at that duration, starting at the first sample of relevant signal.

Any remainder is discarded.

Discontinuous signal concatenation is unavailable with this option.

You may also set an overlap between consecutive segments, using <code class="docutils literal notranslate">Overlap</code> or <code class="docutils literal notranslate">Step</code>.

<code class="docutils literal notranslate">Overlap</code> is expressed as a ratio of duration, between 0 and 1.

<code class="docutils literal notranslate">Step</code> sets the distance in seconds between each consecutive segment.

Chunking <code class="docutils literal notranslate">by longest run</code> gathers all relevant signal cut up into the longest continuous (uninterrupted) segments.

For example, if in the sleep cycle selected there is a run of 10 minutes of REM, with another isolated 30s epoch of REM cut off from the rest, this option will return two segments, one 10 minutes long and one 30s long.

</section>

<h2>Location<a class="headerlink" href="#location" title="Permalink to this heading"></a></h2>

Next, you will see <code class="docutils literal notranslate">Channel group</code> and <code class="docutils literal notranslate">Channel(s)</code>. You may select signal from several channels within a same group.

If you have delimited cycles (see <a class="reference internal" href="notes.html#notes">Annotations</a>), the cycle indices will appear under <code class="docutils literal notranslate">Cycle(s)</code>.

You may select one or several cycles in which to find epochs or events.

If no cycle is selected, data selection will ignore cycles.

You may also select one or several <code class="docutils literal notranslate">Stage(s)</code> in which to find epochs or events.

If no stage is selected, data selection will ignore stages.

</section>

<h2>Rejection<a class="headerlink" href="#rejection" title="Permalink to this heading"></a></h2>

<code class="docutils literal notranslate">Exclude Poor signal epochs</code> removes epochs marked as <code class="docutils literal notranslate">Poor</code> signal.

<code class="docutils literal notranslate">Exclude Artefact events</code> removes signal concurrent with <code class="docutils literal notranslate">Artefact</code> events on all channels.

<code class="docutils literal notranslate">Minimum duration (sec)</code> sets the minimum length of segments after artefact rejection and concatenation, below which the segment is excluded.

For <code class="docutils literal notranslate">Poor</code> signal and <code class="docutils literal notranslate">Artefact</code> event marking, see <a class="reference internal" href="notes.html#notes">Annotations</a>.

</section>

<h2>Concatenation<a class="headerlink" href="#concatenation" title="Permalink to this heading"></a></h2>

You can concatenate different stages, cycles, event types or channels.

Discontinuous signal will only be concatenated if the <code class="docutils literal notranslate">Concatenate discontinuous signal</code> box is checked.

This holds for signal discontinuities introduced by artefact rejection.

Channel concatenation is only available if discontinuous signal is concatenated.

Discontinuous signal concatenation is not available for <code class="docutils literal notranslate">by epoch</code> chunking.

Concatenation is not available when the <code class="docutils literal notranslate">Lock to staging epochs</code> option is selected.

</section>

This box dynamically displays the number of segments relevant to your data selection.

A segment is a slice of signal in time. It may contain data from one or several channels.

Each channel in a segment is analyzed independently, and is represented by one row in the CSV output.

For example, if you have 30 segments over 3 channels, the CSV output will have 90 rows.

</section>

<h2>Pre-processing<a class="headerlink" href="#pre-processing" title="Permalink to this heading"></a></h2>

You may apply any or all of the following transformations to the signal before running the analyses on the right-hand side of the console.

The selected transformations are applied in the order displayed.

<li><code class="docutils literal notranslate">Remove 1/f</code>: Removes the signal’s 1/f background activity, which can help bring out spectral peaks in the signal, above and beyond background activity. In Wonambi’s implementation, this is achieved by subtracting each signal sample from the next sample, effectively resulting in a ‘change’ signal.</li>

<li><code class="docutils literal notranslate">Bandpass</code>: You may apply a variety of bandpass filter types with the drop-down menu, with options for filter <code class="docutils literal notranslate">Order</code>, <code class="docutils literal notranslate">Lowcut</code> and <code class="docutils literal notranslate">Highcut</code>. You may instead choose to only apply a lowpass or highpass filter; in this case, only type in the <code class="docutils literal notranslate">Highcut</code> or <code class="docutils literal notranslate">Lowcut</code>, respectively.</li>

<li><code class="docutils literal notranslate">Notch</code>: You may apply up to two notch filters (a.k.a. powerline filters). To do so, select a filter-type in the drop-down menu and enter the <code class="docutils literal notranslate">Order</code>, <code class="docutils literal notranslate">Centre frequency</code> and <code class="docutils literal notranslate">Bandwidth</code>. Frequencies between (centre frequency +/- (bandwidth / 2)) will be attenuated.</li>

</ul>

</section>

<h2>Frequency<a class="headerlink" href="#frequency" title="Permalink to this heading"></a></h2>

Wonambi offers a highly-customizable range of frequency domain transformations.

For an in-depth discussion of the tools, see Analysis/Frequency Domain.

To activate frequency domain analysis, check <code class="docutils literal notranslate">Compute frequency domain</code>.

To apply the selected pre-processing before the frequency domain analysis, check <code class="docutils literal notranslate">Pre-process</code>.

To obtain a summary spectral plot, averaging all segments and channels, check <code class="docutils literal notranslate">Plot mean spectrum</code>.

To obtain a parametrization of the periodic components of the signal using the FOOOF algorithm (Haller et al., 2018), check <code class="docutils literal notranslate">Parametrize</code>.

Options

This box controls the data export options, as well as pre-processing.

<code class="docutils literal notranslate">Pre-process</code>: if checked, the raw data will be processed according to the options selected in the Pre-processing box, before frequency analyses are applied.

<code class="docutils literal notranslate">Full-spectrum</code>: if checked, the full frequency spectrum will be exported in CSV format, with the suffix ‘_freq.csv’. Rows are segments and columns are sample frequencies from 0 to the Nyquist frequency.

<code class="docutils literal notranslate">Band-limited</code>: if checked, band-limited power will be computed for the bands specified in the Define bands box. results will be exported in CSV format, with the suffix ‘_band.csv’. Rows are segments and columns are bands.

<code class="docutils literal notranslate">Plot mean spectrum</code>: if checked, a summary spectral plot will be displayed, averaging all segments.

<code class="docutils literal notranslate">Parametrize</code>: if checked, the resulting spectrum will be analyzed using the FOOOF algorithm (Haller et al., 2018). Results will be exported to CSV format, with the suffix ‘_fooof.csv’.

Note

The full spectrum, mean spectrum plot and FOOOF parametrization can only be obtained if each transformed segment has the same number of frequency bins, i.e. the same frequency granularity.

Frequency granularity is set by the FFT length, which in a simple periodogram is equal to the segment length.

As a result, it is not possible to obtain the mean of a simple periodogram if the input segments vary in length, as would likely be the case if analyzing events or longest runs.

There are a few workarounds:

<li>Use a <code class="docutils literal notranslate">Time-averaged</code> periodogram, a.k.a. Welch’s method; in this case, FFT length is set by the time window <code class="docutils literal notranslate">Duration</code>. However, time-averaging is impractical for short data segments such as spindles.</li>

<li>Set a <code class="docutils literal notranslate">Fixed</code> FFT length; in this case, shorter segments will be zero-padded to the FFT length, but longer segments will be truncated (not recommended).</li>

<li>Use <code class="docutils literal notranslate">Zero-pad to longest segment</code> to set FFT length to the longest segment and zero-pad all shorter ones. This option is recommended for short data segments such as spindles.</li>

</ol>

</div></blockquote>

</div>

Parameters

<code class="docutils literal notranslate">Scaling</code> sets the type of frequency domain transformation.

To obtain the power spectral density (PSD), set <code class="docutils literal notranslate">Scaling</code> to ‘power’.

For the energy spectral density (ESD), set it to ‘energy’.

The ‘fieldtrip’ and ‘chronux’ type transformations are also provided, but note that these may violate Parseval’s theorem.

<code class="docutils literal notranslate">Taper</code> sets the type of tapering function (a.k.a. windowing function) to use.

Commonly used tapers are ‘boxcar’, ‘hann’ and ‘dpss’ (see below for ‘dpss’).

<code class="docutils literal notranslate">Detrend</code> sets the type of detrending to apply: ‘linear’, ‘constant’ or ‘none’.

If <code class="docutils literal notranslate">Time-averaged</code> is checked, the data will be windowed according to the parameters in the Time averaging box.

Time averaging is used in Bartlett’s method and the closely related Welch’s method.

Time averaging

This box is activated by the <code class="docutils literal notranslate">Time-averaged</code> checkbox in the Parameters box.

It controls the length and spacing of the time windows.

You must set a <code class="docutils literal notranslate">Duration</code>, in seconds, and either an <code class="docutils literal notranslate">Overlap</code> or <code class="docutils literal notranslate">Step</code>.

<code class="docutils literal notranslate">Overlap</code> is expressed as a ratio of Duration, between 0 and 1.

An <code class="docutils literal notranslate">Overlap</code> greater than 0 is equivalent to Welch’s method; at 0 it is equivalent to Bartlett’s method.

Alternatively, you may use <code class="docutils literal notranslate">Step</code> to set the distance in seconds between each consecutive window.

FFT length

This box sets the window length for the Fourier transform.

An FFT length that is <code class="docutils literal notranslate">Same as segment</code> is best for most purposes.

But in cases where you want to, for instance, average the spectra of data segments of varying lengths, you may want to set a fixed FFT length.

To do this, you may either set it manually with <code class="docutils literal notranslate">Fixed</code> or automatically with <code class="docutils literal notranslate">Zero-pad to longest segment</code>.

In the latter case, the FFT length is set to the length of the longest segment N, and zeros are added to the end of all shorter segments until they reach length N.

Zero-padding is a computationally efficient way to effectively interpolate a coarse-grained frequency spectrum to a finer grain.

Multitaper (DPSS) smoothing

This box is activated if ‘dpss’ is selected as <code class="docutils literal notranslate">Taper</code> in the Parameters box.

Here you can set the smoothing parameters for the DPSS/Multitaper method.

<code class="docutils literal notranslate">Half bandwidth</code> sets the frequency smoothing from - half bandwidth to + half bandwidth.

You may normalize the halfbandwidth with <code class="docutils literal notranslate">Normalized</code> (NW = halfbandwidth * duration).

The number of DPSS tapers is then 2 * NW - 1.

Define bands

This box is activated by the <code class="docutils literal notranslate">Band-limited</code> checkbox in Options.

You may enter bands of interest in either list or dynamic notation.

List notation: [[f1,f2],[f3,f4],[f5,f6],…,[fn,fm]]

e.g. [[0.5-4],[4-8],[10-16],[50-100]]

Dynamic notation: (start, stop, width, step)

e.g. (35, 56, 10, 5), equivalent to [[30-40],[35-45],[40-50],[45-55],[50-60]] in list notation.

Note that ‘start’ and ‘stop’ are centre frequencies. Also note that ‘start’ is inclusive, while ‘stop’ is exclusive.

Output

Use this box to select a <code class="docutils literal notranslate">Spectral density</code> output or a <code class="docutils literal notranslate">Complex</code> output.

Normalization

You may normalize the resulting spectral data, either with respect to its own integral or with respect to a normalization period.

When normalizing with respect to a normalization period, the selected frequency analyses are applied directly to the normalization period signal.

To normalize a signal to its integral, select <code class="docutils literal notranslate">by integral of each segment</code> in the drop-down menu.

Each power value will then be divided by the average of all power values for that segment.

To normalize with respect to a normalization period, you must first demarcate this period, either using Event Type(s) or Stage(s).

For example, you may have recorded a quiet wakefulness period at the start of the recording.

In this case, you may create a new Event Type and call it something like ‘qwak’ and mark the entire period as an event on the trace.

You may need to increase the Window Length (in View or on the toolbar) in order to mark the entire period within one window.

Note

In Wonambi, events are channel-specific by default, but for the purposes of demarcating a normalization period, you may mark events on any channel in the channel group.

Just make sure the channel is still in the channel group at the moment of analysis.

</div>

Once the normalization period is marked as one or several ‘qwak’ events, select <code class="docutils literal notranslate">by mean of event type(s)</code> in the drop-down menu and select ‘qwak’ in the <code class="docutils literal notranslate">Event type(s)</code> list.

The power values for each segment will then be divided by the mean power values of all ‘qwak’ events.

Alternatively, you may want to normalize with respect to a stage mean.

In this case, select <code class="docutils literal notranslate">by mean of stage(s)</code> and select the desired stage(s) in the <code class="docutils literal notranslate">Stage(s)</code> list.

The power values for each segment will then be divided by the mean power values for all 30-s epochs of the selected stage(s).

Warning

Normalizing by stage(s) may extend processing time considerably.

</div>

For event type and stage normalization, you may choose to concatenate all relevant normalization periods before applying the frequency transformation, instead of first applying the transformation and then averaging.

To do this, check <code class="docutils literal notranslate">Concatenate</code>.

Note

Like the mean spectral plot, normalization is only available if each segment has the same frequency granularity.

See the note about frequency granularity above.

</div>

Parametrization

Wonambi allows parametrization of power spectra using the FOOOF algorithm:

Haller M, Donoghue T, Peterson E, Varma P, Sebastian P, Gao R, Noto T, Knight RT, Shestyuk A,

Voytek B (2018) Parameterizing Neural Power Spectra. bioRxiv, 299859.

doi: <a class="reference external" href="https://doi.org/10.1101/299859">https://doi.org/10.1101/299859</a>

From the FOOOF Github page:

<div>FOOOF is a fast, efficient, physiologically-informed model to parameterize neural power spectra, characterizing both the 1/f background, and overlying peaks (putative oscillations).

The model conceives of the neural power spectrum as consisting of two distinct functional processes:

<li>A 1/f background, modeled with an exponential fit, with:</li>

<li>Band-limited peaks rising above this background (modeled as Gaussians).</li>

</ul>

</div></blockquote>

With regards to examing peaks in the frequency domain, as putative oscillations, the benefit of the FOOOF approach is that these peaks are characterized in terms of their specific center frequency, amplitude and bandwidth without requiring predefining specific bands of interest.

In particular, it separates these peaks from a dynamic, and independently interesting 1/f background.

</div></blockquote>

If selected, the algorithm will create a CSV report:

You may adjust the following parameters:

<li><code class="docutils literal notranslate">Min. frequency</code> and <code class="docutils literal notranslate">Max. frequency</code>: set the frequency range across which to model the spectrum.</li>

<li><code class="docutils literal notranslate">Peak threshold</code>: sets a threshold above which a peak amplitude must cross to be included in the model. This parameter is in terms of standard deviation above the noise of the flattened spectrum.</li>

<li><code class="docutils literal notranslate">Max. number of peaks</code>: sets the maximum number of peaks to fit (in decreasing order of amplitude).</li>

<li><code class="docutils literal notranslate">Min. peak amplitude</code>: sets an absolute limit on the minimum amplitude (above background) for any extracted peak.</li>

<li><code class="docutils literal notranslate">Min. peak width</code> and <code class="docutils literal notranslate">Max. peak width</code>: set the possible lower- and upper-bounds for the fitted peak widths.</li>

<li><code class="docutils literal notranslate">Background fitting mode</code>: ‘knee’ allows for modelling bends, or knees, in the aperioic signal that are present in broad frequency ranges, especially in intracranial recordings. ‘fixed’ models with a zero knee parameter.</li>

</ul>

</section>

<h2>Phase-amplitude coupling (PAC)<a class="headerlink" href="#phase-amplitude-coupling-pac" title="Permalink to this heading"></a></h2>

Wonambi’s analysis console offers a phase-amplitude coupling analysis (PAC) GUI that ports directly to the <a class="reference external" href="https://github.com/EtienneCmb/tensorpac">Tensorpac</a> package, by Etienne Combrisson.

In order to compute PAC, you must first install tensorpac from the command line (PC) or terminal (Mac):

<code class="docutils literal notranslate">pip install tensorpac</code>

In the analysis console, select <code class="docutils literal notranslate">Compute PAC</code> to enable PAC analysis, and select <code class="docutils literal notranslate">Pre-process</code> to apply the selected pre-processing transformations before analysis.

Choose a <code class="docutils literal notranslate">PAC metric</code> in the drop-down menu.

You may enter one or several phase and amplitude frequencies. Band limits should be separated by a hyphen ‘-’, and each band enclosed in square brackets ‘[]’, separated by commas ‘,’. The entirte epression should also be enclosed in square brackets.

For example, to detect coupling between delta (0.5-4 Hz) and theta (4-8 Hz) as phase frequencies, and low gamma (LG; 30-60 Hz) and high gamma (HG; 60-120 Hz) as amplitude frequencies, you would enter:

<code class="docutils literal notranslate">[[0.5-4],[4-8]]</code> for <code class="docutils literal notranslate">Phase frequencies</code>, and

<code class="docutils literal notranslate">[[30-60],[60-120]]</code> for <code class="docutils literal notranslate">Amplitude frequencies</code>.

This will yield 4 PAC values per segments: delta-LG, delta-HG, theta-LG and theta-HG PAC.

Alternatively, you may use dynamic notation in this format: (start, stop, width, step).

For example, to get the range of amplitude bands between 30 Hz and 130 Hz in non-overlapping 20-Hz bands, you would enter:

<code class="docutils literal notranslate">(40,140,20,20)</code>.

Notice that start and stop are centre frequencies. Notice also that start is inclusive but stop is exclusive, so in order to capture 110-130 Hz, stop must be set after the centre frequency, i.e. 121-140.

For more information, see the <a class="reference external" href="https://etiennecmb.github.io/tensorpac/">Tensorpac documentation</a>.

</section>

<h2>Events<a class="headerlink" href="#events" title="Permalink to this heading"></a></h2>

The console’s Events tab allows the extraction of a suite of commonly studied parameters.

Event parameters are divided into global parameters, local parameters and slow wave parameters.

Global

<li><code class="docutils literal notranslate">Count</code> simply returns the number of segments.</li>

<li><code class="docutils literal notranslate">Density, per</code> returns the number of segments divided by the number of epochs of relevant signal. The relevant signal is all epochs corresponding to the cycle(s) and stage(s) selected in the Location box. You may set the epoch length in seconds with the text box.</li>

</ul>

Band of interest

For <code class="docutils literal notranslate">Power</code>, <code class="docutils literal notranslate">Energy</code>, <code class="docutils literal notranslate">Peak power frequency</code> and <code class="docutils literal notranslate">Peak energy frequency</code>, you may set a band of interest.

These analyses are then carried out only over that spectral band.

If no frequencies are specified, analyses are applied to the entire spectrum.

Local

For each parameter, check the box next to it to extract it, and select the corresponding box in the <code class="docutils literal notranslate">Pre-process</code> column in order to apply the selected pre-processing before analysis.

Note that for all parameters except <code class="docutils literal notranslate">Duration</code>, the output will contain one value per channel per segment.

<li><code class="docutils literal notranslate">Duration</code>: The segment duration, in seconds.</li>

<li><code class="docutils literal notranslate">Min. amplitude</code>: The lowest amplitude value in the signal.</li>

<li><code class="docutils literal notranslate">Max. amplitude</code>: The highest amplitude value in the signal.</li>

<li><code class="docutils literal notranslate">Peak-to-peak amplitude</code>: The absolute difference between the lowest and highest amplitude values in the signal.</li>

<li><code class="docutils literal notranslate">RMS</code>: The square root of the mean of the squares of each amplitude value in the signal.</li>

<li><code class="docutils literal notranslate">Power</code>: The average power (from a simple periodogram) of the signal over the band of interest. Best used for stationary signals.</li>

<li><code class="docutils literal notranslate">Energy</code>: The average energy (from a simple periodogram) of the signal over the band of interest. Best used for signals with a clear beginning and end, i.e. events.</li>

<li><code class="docutils literal notranslate">Peak power frequency</code>: The frequency corresponding to the highest power value in the band of interest.</li>

<li><code class="docutils literal notranslate">Peak energy frequency</code>: The frequency corresponding to the highest energy value in the band of interest.</li>

</ul>

Slow wave

These are local parameters that apply only to slow waves.

You may still apply these analyses to any signal, but if the signal does not have the morphological characteristics of a slow wave, the output will be nan (not a number).

<code class="docutils literal notranslate">Average slopes</code> and <code class="docutils literal notranslate">Max. slopes</code> each return 5 values: one per slow wave quadrant and a fifth for the combination of quadrants 2 and 3:

<li>Q1: First zero-crossing to negative trough</li>

<li>Q2: Negative trough to second zero-crossing</li>

<li>Q3: Second zero-crossing to positive peak</li>

<li>Q4: Positive peak to third zero-crossing</li>

<li>Q23: Negative trough to positive peak.</li>

</ul>

<code class="docutils literal notranslate">Average slopes</code> is the amplitude difference between the quadrant start and end divided by the quadrant duration, in μV/s.

<code class="docutils literal notranslate">Max. slopes</code> is the maximum value of the derivative of the smoothed signal (50-ms moving average) of the quadrant, in μV/s2.

</section>

</div>

<a href="methods.html" class="btn btn-neutral float-left" title="Detection methods" accesskey="p" rel="prev"> Previous</a>

</div>

<hr/>

</div>

</footer>

</div>

</section>

</div>

jQuery(function () {

SphinxRtdTheme.Navigation.enable(false);

});

</script>

</body>

</html>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

FilesExpand file tree

analysis.html

Latest commit

History

analysis.html

File metadata and controls