[MRG] DOC: Update alpha example #249

ntolley · 2021-01-20T23:09:49Z

Adding more detailed descriptions to alpha wave example. Additionally updating simulation to match alpha and beta tutorial provided in HNN-GUI (https://jonescompneurolab.github.io/hnn-tutorials/alpha_and_beta/alpha_and_beta)

jasmainak · 2021-01-20T23:57:11Z

Let me know when it's ready!

examples/plot_simulate_alpha.py

jasmainak · 2021-01-25T04:21:43Z

you need to rebase

also the CircleCI error can be resolved in many ways

Increasing the CircleCI timeout
Changing the sphinx building to be more verbose (currently it suppresses example outputs)
Separating the example into two

Choose your poison but I'd prefer 2. or 3.

cjayb · 2021-01-25T04:35:22Z

What's the reason for increasing the simulation time from 310 to 710? I get it from a physiological perspective, but to keep the examples running in reasonable time, I'd prefer we stuck to the max 300 we have in all other cases.

Doing low-F TFRs gets tricky, so we should consider not including too many of them. The spectrum isn't as pretty, but how about just using that in this example?

cjayb

As mentioned in the comments, really think this should be a 300 ms simulation. Even at the cost of a pretty TFR...

examples/plot_simulate_alpha.py

cjayb · 2021-01-25T06:15:19Z

examples/plot_simulate_alpha.py

+syn_delays_p = {'L2_basket': 0.1, 'L2_pyramidal': 0.1,
+                'L5_basket': 1., 'L5_pyramidal': 1.}


Suggested change

syn_delays_p = {'L2_basket': 0.1, 'L2_pyramidal': 0.1,

'L5_basket': 1., 'L5_pyramidal': 1.}

syn_delays_p = {'L2_pyramidal': 0.1, 'L5_pyramidal': 1.}

cjayb · 2021-01-25T06:16:22Z

examples/plot_simulate_alpha.py

+syn_delays_d = {'L2_basket': 5., 'L2_pyramidal': 5.,
+                'L5_basket': 5., 'L5_pyramidal': 5.}


Suggested change

syn_delays_d = {'L2_basket': 5., 'L2_pyramidal': 5.,

'L5_basket': 5., 'L5_pyramidal': 5.}

syn_delays_d = {'L2_pyramidal': 5., 'L5_pyramidal': 5.}

Are these default values for basket cells or something? In any case I am getting this error when I try it:
ValueError: synaptic_delays is either a common float or needs to be specified as a dict for each cell type

Nope: the delays need to be specified for any cell type for which weights are given. Since you're only connecting to pyramidal cells, there's no reason to include basket delays in the dict.

This works for me on this branch, could you try again before we investigate further?

examples/plot_simulate_alpha.py

ntolley · 2021-01-25T14:52:48Z

I'll see how everything looks with the 300 ms simulation. The main motivation to increase the timing was to match the alpha/beta tutorial for HNN-GUI. It is necessary because the only beta event in the simulation occurs after 300 ms: https://jonescompneurolab.github.io/hnn-tutorials/alpha_and_beta/alpha_and_beta.

Since the timing of beta events is stochastic, I would need to find a seed that produces a beta event within 0-300 ms. I'm not against this, but it's just another question of how much congruency we want to maintain with the GUI tutorials.

ntolley · 2021-01-25T14:57:46Z

Choose your poison but I'd prefer 2. or 3.

@jasmainak I think the path will just be decreasing the simulation time and changing the seeding to produce beta events earlier in the simulation. I started making a separate example for just beta events but it seemed pretty redundant as I got working. The only distinction from the alpha example is a second 10 Hz rhythmic drive.

cjayb · 2021-01-25T16:53:49Z

I think the path will just be decreasing the simulation time and changing the seeding to produce beta events earlier in the simulation

I agree! It should be possible to /infer/ the seed need to produce the ~450 ms event, but it's probably easier to go with trial-and-error ;)

jasmainak · 2021-01-25T19:59:16Z

examples/plot_simulate_alpha.py

 params_fname = op.join(hnn_core_root, 'param', 'default.json')
 params = read_params(params_fname)
 print(params)

 ###############################################################################
-# Update a few of the default parameters related to visualisation
+# Next we update a few of the default parameters related to visualisation. The
+# ``dipole_scalefctr`` relates to the amount of cortical tissue necessary to


I wonder if we should instead add this line at the end:

simulate_dipole(..., postproc=False) dpl.scale(15000)

to discourage fiddling with params as much as possible.

This also makes me realize that Dipole is not documented in the API reference. Do you mind adding it there?

+1!

I’d go as far as to stop reading any post-proc details from params (got bit in the rear by a dipole_smooth_win setting of 30, set to zero above). future PR...

Good suggestion! I've removed all params edits except for tstop now,

examples/plot_simulate_alpha.py

jasmainak · 2021-01-25T23:11:18Z

@ntolley thanks for addressing all our whims!

so the beta rhythm doesn't really look rhythmic to me and the previous alpha rhythm was more rhythmic. Does it look rhythmic when you plot until 710 ms? Spectrogram plots can be deceptive as even a small non-stationarity in the signal can show up as a rhythm. I think I prefer TFR for this reason as you can see clearly then how long is the beta -- whether it's something long enough to be considered a brain rhythm.

Many people also use Welch instead of regular PSD ... and you'll see this peak disappear if you use Welch.

Not to question everything but perhaps something to keep in mind? Thoughts?

ntolley · 2021-01-25T23:41:21Z

@jasmainak it's a valid point, the beta rhythm isn't really a rhythm which is highly emphasized in the references/GUI tutorial. At the same time I don't think this negates the utility of power spectra for identifying discrete beta events (in fact they are present in the GUI tutorial). If we really want to get into the nuance of the phenomena and how it's measured this example would need to be a bit longer.

Also it's not exactly clear in the TFR that these beta events are in fact discrete events when using shorter simulations. In light of at least showing beta band activity they serve the same purpose (although I'll need to check on the welch analysis...)

If you feel strongly for the TFR plots I can change over the plots.

codecov-io · 2021-01-26T00:42:08Z

Codecov Report

Merging #249 (7fd2944) into master (13c2dfe) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #249   +/-   ##
=======================================
  Coverage   52.93%   52.93%           
=======================================
  Files          23       23           
  Lines        2747     2747           
=======================================
  Hits         1454     1454           
  Misses       1293     1293

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 13c2dfe...7fd2944. Read the comment docs.

jasmainak · 2021-01-26T00:46:32Z

okay all good. Last question -- are we okay to have the alpha rhythm deviate from the example on the HNN website?

ntolley · 2021-01-26T01:08:09Z

@rythorpe what do you think? My feeling is it's ok to deviate if we're already truncating the recording to half the original length. But it's a pretty simple fix to change the seeding back. I'll just have to play with the seeding of the second inputs that are added to ensure a beta event occurs within the time window.

jasmainak · 2021-01-26T02:29:23Z

well this example was kind of a visual check for the rhythmic drive since we don't have a proper test for it. My eye had gotten used to spotting any discrepancies. But I won't insist on keeping the old waveforms if you guys are okay with changing. I let @rythorpe merge if he's happy!

cjayb

Many people also use Welch instead of regular PSD ... and you'll see this peak disappear if you use Welch.

Good point. The current implementation uses spectrogram with n_fft set to the entire window. I think a more principled approach to both spectra and tfr's are called for. I think we should optimise the examples for aesthetics.

examples/plot_simulate_alpha.py

cjayb · 2021-01-26T14:15:52Z

Looking good! I tried to fiddle with different TFR settings, but it's hard to make anything illustrative with only one trial (which is very much to the point!). The spectra do the trick of underlining the point.

Take a look at Figure 4 in Sherman et al. (2016) below, which hnn_core can clearly reproduce:

The beta-events appear because the proximal drive has a wider burst_std compared with the distal one.

@ntolley to 'guarantee' beta event, you could reduce burst_std to 15 or even 10 ms for the second distal drive. I would also modify the comment/explanation in the code to something like:

# The next step is to add a simultaneous 10 Hz distal drive with a lower
# within-burst spread of spike times (``burst_std``) compared with the
# proximal one. The different arrival times of spikes at opposite ends of
# the pyramidal cells will tend to produce bursts of 15-30 Hz power known
# as beta frequency events.

Co-authored-by: Christopher J. Bailey <bailey.cj@gmail.com>

jasmainak · 2021-01-27T05:13:01Z

merge button is yours @cjayb ! All good on my end

cjayb · 2021-01-27T14:26:12Z

Hurra! See follow-up at #264

ntolley force-pushed the alpha_docs branch from 10f2b2b to 6f4af05 Compare January 23, 2021 00:23

cjayb reviewed Jan 23, 2021

View reviewed changes

examples/plot_simulate_alpha.py Outdated Show resolved Hide resolved

cjayb reviewed Jan 23, 2021

View reviewed changes

examples/plot_simulate_alpha.py Outdated Show resolved Hide resolved

cjayb reviewed Jan 23, 2021

View reviewed changes

examples/plot_simulate_alpha.py Show resolved Hide resolved

cjayb reviewed Jan 23, 2021

View reviewed changes

examples/plot_simulate_alpha.py Outdated Show resolved Hide resolved

cjayb approved these changes Jan 25, 2021

View reviewed changes

jasmainak reviewed Jan 25, 2021

View reviewed changes

examples/plot_simulate_alpha.py Outdated Show resolved Hide resolved

ntolley force-pushed the alpha_docs branch from dc563df to 5d42554 Compare January 25, 2021 22:56

ntolley changed the title ~~DOC: Update alpha example~~ [MRG] DOC: Update alpha example Jan 25, 2021

jasmainak approved these changes Jan 26, 2021

View reviewed changes

cjayb reviewed Jan 26, 2021

View reviewed changes

examples/plot_simulate_alpha.py Outdated Show resolved Hide resolved

examples/plot_simulate_alpha.py Outdated Show resolved Hide resolved

examples/plot_simulate_alpha.py Show resolved Hide resolved

ntolley and others added 5 commits January 26, 2021 13:39

Add more commentary to the example docs

3c422dc

Add proximal drive to alpha example to simulate beta events

1931172

Remove redundant figure

b70f9da

Fix plotting and descriptions, add references

b4074e0

Update examples/plot_simulate_alpha.py

83467c0

Co-authored-by: Christopher J. Bailey <bailey.cj@gmail.com>

ntolley and others added 11 commits January 26, 2021 13:39

Update examples/plot_simulate_alpha.py

74d033a

Co-authored-by: Christopher J. Bailey <bailey.cj@gmail.com>

Update examples/plot_simulate_alpha.py

1cab747

Co-authored-by: Christopher J. Bailey <bailey.cj@gmail.com>

Update examples/plot_simulate_alpha.py

9d69570

Co-authored-by: Christopher J. Bailey <bailey.cj@gmail.com>

Update examples/plot_simulate_alpha.py

0b5a2d6

Co-authored-by: Christopher J. Bailey <bailey.cj@gmail.com>

Update examples/plot_simulate_alpha.py

8f4c73e

Co-authored-by: Christopher J. Bailey <bailey.cj@gmail.com>

Improved simulation seeding for beta events and docs updates

cea9a54

Replace TFR plot with PSD, improved descriptions

05af0c1

Prettier plots

6ad0be9

Flake8

44c31ee

Update api.rst

99fa860

Improve DOC descriptions, remove dipole scaling

c787ff7

ntolley force-pushed the alpha_docs branch from 7f9bbe6 to c787ff7 Compare January 26, 2021 18:39

Reduce distal std

c89022b

cjayb merged commit acef0cd into jonescompneurolab:master Jan 27, 2021

cjayb added a commit to cjayb/hnn-core that referenced this pull request Jan 27, 2021

MAINT small cosmits after jonescompneurolab#249

56bd90c

cjayb mentioned this pull request Jan 27, 2021

[MRG] updated visualisation methods and unit handling #264

Merged

7 tasks

cjayb added a commit to cjayb/hnn-core that referenced this pull request Feb 2, 2021

MAINT small cosmits after jonescompneurolab#249

f049bff

jasmainak pushed a commit that referenced this pull request Feb 25, 2021

MAINT small cosmits after #249

663f56b

ntolley mentioned this pull request Feb 16, 2022

Bestpractice #464

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[MRG] DOC: Update alpha example #249

[MRG] DOC: Update alpha example #249

ntolley commented Jan 20, 2021

jasmainak commented Jan 20, 2021

jasmainak commented Jan 25, 2021 •

edited

Loading

cjayb commented Jan 25, 2021

cjayb left a comment

cjayb Jan 25, 2021

cjayb Jan 25, 2021

ntolley Jan 25, 2021

cjayb Jan 26, 2021

ntolley commented Jan 25, 2021

ntolley commented Jan 25, 2021

cjayb commented Jan 25, 2021

jasmainak Jan 25, 2021

cjayb Jan 25, 2021

ntolley Jan 25, 2021

jasmainak commented Jan 25, 2021 •

edited

Loading

ntolley commented Jan 25, 2021

codecov-io commented Jan 26, 2021 •

edited

Loading

jasmainak commented Jan 26, 2021

ntolley commented Jan 26, 2021 •

edited

Loading

jasmainak commented Jan 26, 2021

cjayb left a comment

cjayb commented Jan 26, 2021

jasmainak commented Jan 27, 2021

cjayb commented Jan 27, 2021

		syn_delays_p = {'L2_basket': 0.1, 'L2_pyramidal': 0.1,
		'L5_basket': 1., 'L5_pyramidal': 1.}

	syn_delays_p = {'L2_basket': 0.1, 'L2_pyramidal': 0.1,
	'L5_basket': 1., 'L5_pyramidal': 1.}
	syn_delays_p = {'L2_pyramidal': 0.1, 'L5_pyramidal': 1.}

		syn_delays_d = {'L2_basket': 5., 'L2_pyramidal': 5.,
		'L5_basket': 5., 'L5_pyramidal': 5.}

	syn_delays_d = {'L2_basket': 5., 'L2_pyramidal': 5.,
	'L5_basket': 5., 'L5_pyramidal': 5.}
	syn_delays_d = {'L2_pyramidal': 5., 'L5_pyramidal': 5.}

[MRG] DOC: Update alpha example #249

[MRG] DOC: Update alpha example #249

Conversation

ntolley commented Jan 20, 2021

jasmainak commented Jan 20, 2021

jasmainak commented Jan 25, 2021 • edited Loading

cjayb commented Jan 25, 2021

cjayb left a comment

Choose a reason for hiding this comment

cjayb Jan 25, 2021

Choose a reason for hiding this comment

cjayb Jan 25, 2021

Choose a reason for hiding this comment

ntolley Jan 25, 2021

Choose a reason for hiding this comment

cjayb Jan 26, 2021

Choose a reason for hiding this comment

ntolley commented Jan 25, 2021

ntolley commented Jan 25, 2021

cjayb commented Jan 25, 2021

jasmainak Jan 25, 2021

Choose a reason for hiding this comment

cjayb Jan 25, 2021

Choose a reason for hiding this comment

ntolley Jan 25, 2021

Choose a reason for hiding this comment

jasmainak commented Jan 25, 2021 • edited Loading

ntolley commented Jan 25, 2021

codecov-io commented Jan 26, 2021 • edited Loading

Codecov Report

jasmainak commented Jan 26, 2021

ntolley commented Jan 26, 2021 • edited Loading

jasmainak commented Jan 26, 2021

cjayb left a comment

Choose a reason for hiding this comment

cjayb commented Jan 26, 2021

jasmainak commented Jan 27, 2021

cjayb commented Jan 27, 2021

jasmainak commented Jan 25, 2021 •

edited

Loading

jasmainak commented Jan 25, 2021 •

edited

Loading

codecov-io commented Jan 26, 2021 •

edited

Loading

ntolley commented Jan 26, 2021 •

edited

Loading