How to inspect time-frequency results

preview_player
Показать описание
If you are unsure of how to look at time-frequency results, this video has the 5-step plan that you need! It also discusses whether time-frequency features can be interpreted as "oscillations."

Рекомендации по теме
Комментарии
Автор

I understand about the signal annotation by watching your lecture. Thank you so much. Now I am trying research for EEG signal feature annotation with machine learning approach.

zinwaroo
Автор

Hi, me again...the one who asked about R. Then I go to matlab...to compare... I used cwt to transform a monthly precipitation data. I have 1320 month. So I set my fs as 1 (month) dt as 1, time start as 0 time end as 1320. Then I use [wt, f, coi] =cwt(precipitation, ‘amor’, fs). Then my y-axis is frequency. The unit shows mhz. I am confuse here...how the wavelet window’s frequency connects to the time of the time series? See vertically I see frequency range from 0 to 100+, so for example what does 10 represent? So can I say this signal occurs each 10 month when there is high magnitude around frequency 10? Hope I made myself clear.

DL-vdmr
Автор

For my monthly data, Is frequency 10 in Y axis mean 10 month? Can I explain it as every 10 month there is wavelength repeated it self if I see a high magnitude.

DL-vdmr
Автор

Hi :) Thank you for that awesome video. I have some trouble to understand the values of the color bar. How do they reflect power? When you do the coding to create such a plot, you set a minimum and maximum value for this color bar. What changes when you change those range? If you set it for example to -0.5 to 0.5 and a value at a given frequency and time is 0.6, is that value not displayed anymore?

franziskal
Автор

Hi Mike, thanks so much for this! When I’m looking at components, instead of channels, at how many components should I look like? I’m unsure between the difference of component and channel (channel is the number of electrode isn’t?)

sophies
Автор

Ha I found it I should use [wt, periods, coi] = cwt(data, ‘amor’, years(1/12))...then Y axis related to time too. Happy about it. Another question do you know how to get significant line. I saw some paper they made significant contours or curves for the magnitude. But don’t know how...

DL-vdmr
Автор

Thank you Mike. I would recommend to include the signal in time domain with time-frequency spectrum in order to have better understanding of the physical property of the system under study.

khalfanalkharusi
Автор

How about extracting meaningful Information from those nice color maps, using Principal Component Analysis (PCA) for instance ?

onusbaum
Автор

Hi Mike,
Thank you for the very informative lectures. I have a very general question. After conducting an experiment under two conditions, collecting data and pre-preprocessing it (including ICA rejection). From where to start the analysis? ERP or TF analysis? Which electrodes and which frequency bands? I am not looking for a specific answer but more for a methodology. Sorry if my question is too basic.

haneensuradi
Автор

What program are you using? Can you do a tutorial ??

stefanyleonpalacios
Автор

Hey Mike. I've commented on one of your other videos before. Learned a lot from them, thanks! It would be cool if you would check out my YouTube channel; I have some videos on additive synthesis and signal recognition. It would be cool if you could take a look! Also, what program do you use for screen capture?-Ryan

rblack
Автор

hii mike, your video is quite helpful. i wanted to send my query to you personally. can you please provide my your email id

jayagupta
visit shbcf.ru