US 11,766,210 B2
Methods and devices for determining signal quality for a swallowing impairment classification model
Michael Reuben Jedwab, Lausanne (CH); Juha M. Kortelainen, Tampere (FI); Rajat Mukherjee, Cambridge, MA (US); and Harri Polonen, Tampere (FI)
Assigned to Societe des Produits Nestle S.A., Vevey (CH)
Appl. No. 16/636,449
Filed by SOCIETE DES PRODUITS NESTLE S.A., Vevey (CH)
PCT Filed Jul. 31, 2018, PCT No. PCT/EP2018/070649
§ 371(c)(1), (2) Date Feb. 4, 2020,
PCT Pub. No. WO2019/030046, PCT Pub. Date Feb. 14, 2019.
Claims priority of provisional application 62/541,943, filed on Aug. 7, 2017.
Prior Publication US 2020/0170562 A1, Jun. 4, 2020
Int. Cl. A61B 5/00 (2006.01); G06F 3/01 (2006.01); G06F 17/18 (2006.01)
CPC A61B 5/4205 (2013.01) [A61B 5/7203 (2013.01); A61B 5/7264 (2013.01); A61B 5/7435 (2013.01); G06F 3/011 (2013.01); G06F 3/017 (2013.01); G06F 17/18 (2013.01); A61B 2562/0219 (2013.01)] 17 Claims
OG exemplary drawing
 
1. A device for screening swallowing safety and swallowing efficiency, the device comprising:
a processor configured to receive accelerometry data of throat vibrations, determine vibration signals of the accelerometry data along an anterior-posterior axis and a superior-inferior axis of the throat to be used in determining a spectrogram, and perform at least one method comprising (i) determining a signal variance of the accelerometry data as a function of time, as summed power over a specific frequency range of the spectrogram, and comparing the signal variance to a first threshold value,
the at least one method further comprising at least one additional method selected from the group consisting of (ii) determining a normalized variance signal from the spectrogram by setting a span between 0 and 1, selecting a beginning portion of the normalized variance signal, and comparing values of the beginning portion of the normalized variance signal to a second threshold value, (iii) determining a normalized variance signal from the spectrogram by setting a span between 0 and 1, selecting an end portion of the normalized variance signal, and comparing values of the end portion of the normalized variance signal to a third threshold value, and (iv) determining a summed power spectral density of both the vibration signals along the anterior-posterior axis and the superior-inferior axis of the throat as an average of the spectrogram over a full duration or signal length relating to consumption of a single bolus and applying spectral entropy for the summed power spectral density (PSD) of the vibration signals along both the anterior-posterior axis and the superior-inferior axis of the throat for comparison to a fourth threshold value, wherein the at least one method is performed by the processor in real-time relative to receipt of the accelerometry data; and
a user interface configured to provide one or more outputs comprising at least one of audio or graphics, and the one or more outputs indicate a missing swallow for the accelerometry data based on one or more results of the at least one method.