System and method for evaluating feeding maturation

1 . A method for monitoring feeding maturation in a premature baby, the method comprising:

measuring an acoustic response from a baby during a selected time period to provide energy constrained spectral content information;

comparing the measured acoustic response against a train data set to determine an indication of a swallow event;

measuring a respiration pattern of the baby during a swallow cycle using a peak and valley model to provide respiration data; and

determining a feeding maturity of the baby in dependence on the swallow indication and respiration data,

wherein measuring the acoustic response from the baby during the selected time period comprises extracting features from a received signal and comparing the extracted features against the train data set, wherein the extracted features may be classified as swallow or non-swallow events, and

wherein, prior to extracting features, a received signal is framed so as to be considered as stationary.

2 . The method of claim 1 , wherein the comparison against the train data set comprises comparing extracted features in a classification process.

3 . The method of claim 2 , wherein the extracted features are determined based on feature extraction performed using a speech analysis tool.

4 . The method of claim 3 , wherein the feature extraction comprises labelling identified features as one or more of features selected from a group including:

a swallow sound;

a final discrete sound;

a respiration sound; and

other non-swallow sounds including a vowel sound, a pleasure sound, or a crying sound.

5 . The method of claim 1 , further comprising determining the feeding maturity based on maximum rhythmic swallow numbers and an average time between swallows.

6 . The method of claim 1 , wherein training of the train data set comprises:

receiving audio samples from healthy subjects;

framing the received audio samples so as to provide stationary signals; and

extracting features from the framed samples so as to generate the train data set.

7 . The method of claim 1 , wherein the peak and valley model to determine respiration data comprises:

receiving an input signal indicating respiration of the baby;

processing the received input signal with a low pass filter so as to provide a smoothed respiration signal; and

extracting parameters associated with the respiration based on the smoothed respiration signal.

8 . The method of claim 7 , wherein the parameters include one or more of a breath rate, and an onset or end of an inspiration or expiration event.

9 . The method of claim 1 , further comprising:

determining an inspiration after swallow count in dependence on the swallow indication and respiration data, wherein the inspiration after swallow count indicates a number of inspiration events occurring just after a swallow event has finished; and

determining that the feeding maturity of the baby has decreased based on the inspiration after swallow count increasing.

10 . A device for monitoring feeding maturation in a premature baby, the device comprising:

a first sensor configured to measure an acoustic response from a baby during a selected time period to provide energy constrained spectral content information;

a second sensor configured to measure a respiration pattern of the baby during a swallow cycle;

a processor coupled to the first and second sensors, the processor configured to:

compare the measured acoustic response against a train data set to determine an indication of a swallow event; and

provide respiration data based on the measured respiration pattern and using a peak and valley model; and

determining a feeding maturity of the baby in dependence on the swallow indication and respiration data,

wherein measuring the acoustic response from the baby during the selected time period comprises extracting features from a received signal and comparing the extracted features against the train data set, wherein the extracted features may be classified as swallow or non-swallow events, and

wherein, prior to extracting features, a received signal is framed so as to be considered as stationary.

11 . The device of claim 10 , wherein the processor is configured to train the data set by:

receiving audio samples from healthy subjects;

framing the received audio samples so as to provide stationary signals; and

extracting features from the framed samples so as to generate the train data set.

12 . The device of claim 10 , wherein the processor is configured to apply the peak and valley model to determine respiration data by:

receiving an input signal indicating respiration of the baby;

processing the received input signal with a low pass filter so as to provide a smoothed respiration signal; and

extracting parameters associated with the respiration based on the smoothed respiration signal.