[FieldTrip] mtmconvol with hanning window (Jochem Rieger)

[Schoffelen, J.M. (Jan Mathijs)]

Dear Dan,

Indeed, it is a bit confusing that the ‘mtmconvol’ method in FieldTrip also allows for single tapers, e.g. the Hanning taper.

Once you use cfg.method = ‘mtmconvol’ (or ‘mtmfft’), and NOT use cfg.taper = ’dpss’/’sine’ then no multitapers will be applied (and also the cfg.tapsmofrq parameter will become irrelevant).

Perhaps this could be documented a bit better, e.g. in the function’s help, and on an appropriate location on the website. So, if you feel adventurous, we are open to suggestions.

In your case, I would describe the part of the methods where you used the Hanning taper for a time-frequency decomposition something like: we applied a sliding time window based Fourier transform to the data, using a Hanning taper to control for spectral leakage.

Best wishes,
Jan-Mathijs




> On 4 Aug 2023, at 18:08, Weisholtz, Daniel S.,MD via fieldtrip <fieldtrip at science.ru.nl> wrote:
>
> Dear Jochem,
>
> Thank you for this detailed discussion.  I wasn’t referring the Welch’s method, because I understand that method collapses across the time dimension and produces a power spectrum. I am interested in the other methods that produce time-frequency maps.
>
> My understanding is that when using mtmconvol, the multitaper method is only applied if the Slepian windows (dpss) are used in the taper field.  With the hanning window, it is a single taper that is applied to overlapping segments of data (that’s what I meant by sliding window).
>
> I do get the advantages and disadvantages of using a rectangular window.  In Mike X. Cohen’s textbook, “Analyzing Neural Time Series Data,” he has a chapter on Short-Time FFT where he recommends using an overlapping taper such as a Hanning window.  So, at least according to him, the STFFT doesn’t need to use strictly rectangular windows.
>
> In any case, how would you describe an analysis done with mtmconvol and a hanning window in the methods section of a paper, if not short-time fft? It seems to me that it would be incorrect to refer to it as a multitaper method.
>
> Thanks again.
>
> Best,
> Dan
>
>>
>> ------------------------------
>>
>> Message: 2
>> Date: Mon, 31 Jul 2023 16:32:28 +0200
>> From: Jochem Rieger <jochem.rieger at uni-oldenburg.de>
>> To: "Weisholtz, Daniel S.,MD via fieldtrip" <fieldtrip at science.ru.nl>
>> Subject: Re: [FieldTrip] mtmconvol with hanning window
>> Message-ID: <596f1ab4-03b0-40c2-92c8-a37597aac43a at uni-oldenburg.de>
>> Content-Type: text/plain; charset="windows-1252"; Format="flowed"
>>
>> Hi Dan,
>>
>> think of short term fft as windowing with a rectangular window. This has
>> rather bad frequency resolution (wide central lobe, many side lobes in
>> the amplitude spectrum). It produces "veridical" values for the
>> amplitude estimate over the analysis  interval (your window length)
>>
>> The other methods underestimate the amplitudes systematically. If you
>> are only interested in relative amplitudes (most cases) that doesn't
>> matter.
>>
>> The Hanning window improves the cross talk and is typically applied
>> once  the whole intervall and then an FFT is calculated to get the
>> amplitude spectrum. The cost for that improvement is that the amplitude
>> estimate are biased towards the center of the window and the amplitude
>> values are underestimated.
>>
>> Multitaper analysis applies multiple orthogonal windows and obtains by
>> that approach multilple amplitude spectra, which are then combined for
>> the final estimate. Approach reduces the noise in the amplitude
>> spectrum  lets  you trade-off frequency resolution against noise.
>>
>> When you write multiple sliding windows you likely refer to the Welch
>> method which applies multiple shorter and partly overlapping tapers to
>> the analysis interval and then combines those correlated amplitude
>> estimates. This method reduces the noise in the amplitude spectrum. In
>> principle, you could use different types of windows.
>>
>> Hope that helps,
>>
>>    Jochem
>>
>>
>> On 27.07.23 18:43, Weisholtz, Daniel S.,MD via fieldtrip wrote:
>>>
>>> ACHTUNG!Diese E-Mail kommt von Extern! WARNING! This email originated
>>> off-campus.
>>>
>>> Dear FieldTrip community,
>>>
>>> I understand that using the hanning window with this function will
>>> make this function use a single sliding taper window rather than
>>> multitaper. Does that make this analysis equivalent to a short-time
>>> fft analysis? If you do this, will FieldTrip ignore the tapsmofrq
>>> parameter?
>>>
>>> Thanks for your help!
>>>
>>> Best,
>>> Dan
>>>
>
> The information in this e-mail is intended only for the person to whom it is addressed.  If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Mass General Brigham Compliance HelpLine at https://urldefense.com/v3/__https://www.massgeneralbrigham.org/complianceline__;!!HJOPV4FYYWzcc1jazlU!_l9kNlAf1yrU7ED9aTxm5aSEzRmz3sU_9xMroXclAIPQXULqhzyz3tmT3E1bNy5rmFJmQYev44qvGAKR6MNSQ3S5jV2Aq-d8$  <https://urldefense.com/v3/__https://www.massgeneralbrigham.org/complianceline__;!!HJOPV4FYYWzcc1jazlU!_l9kNlAf1yrU7ED9aTxm5aSEzRmz3sU_9xMroXclAIPQXULqhzyz3tmT3E1bNy5rmFJmQYev44qvGAKR6MNSQ3S5jV2Aq-d8$ > .
> Please note that this e-mail is not secure (encrypted).  If you do not wish to continue communication over unencrypted e-mail, please notify the sender of this message immediately.  Continuing to send or respond to e-mail after receiving this message means you understand and accept this risk and wish to continue to communicate over unencrypted e-mail.
>
> _______________________________________________
> fieldtrip mailing list
> https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
> https://urldefense.com/v3/__https://doi.org/10.1371/journal.pcbi.1002202__;!!HJOPV4FYYWzcc1jazlU!5ZmiNMJTP4TV0bhP7RO7GMcQwOuWDyAe1lB77gr-uyag2f4LDvl0xbZyyz0rgLYbO6boqR7j0Ch0QoawgnEd20hLz9LiFLWIh6iQ9g$