antonior92 / ecg-age-prediction Goto Github PK

Hi, thank you for sharing the amazing research!

I tried to reproduce the result of fig 2 (Adjusted survival curves) in paper with exam.csv in CODE-15% cohort (https://zenodo.org/record/4916206#.Yib1gRPP30o)

I calculated "diff" columns by

total_data = pd.read_csv('exams_CODE15.csv')
total_data['diff'] = total_data['nn_predicted_age'] - total_data['age']

And I separate the dataset into 3 groups as i1, i2, and i3. I1 was the group that nn_predicted_age is 8 years greater than age, I2 was the one that nn_predicted_age is 8 years smaller than age, I3 was the one within a range of 8 years from their age

from lifelines import KaplanMeierFitter
i1 = (total_data['diff'] > 8)      
i2 = (total_data['diff'] < -8)

i3 = i1+i2
i3 = ~i3
 
durations = total_data['timey']
event_observed = total_data['death'].astype('int')
kmf_old = KaplanMeierFitter() 
kmf_young = KaplanMeierFitter() 

kmf = KaplanMeierFitter() 

kmf_old.fit(durations[i1], event_observed[i1],label='ECG_age_Old_abs')
a1 = kmf_old.plot()

kmf_young.fit(durations[i2], event_observed[i2],label='ECG_age_Young_abs')
a2 = kmf_young.plot()

kmf.fit(durations[i3], event_observed[i3],label='ECG_age_similar_abs')

kmf.plot(ax=a1) 

However, the result that I got from the above code is opposite from your result of the paper.

Can you check the exam.csv in CODE-15% cohort has all right value?

Thank you for reading
Best Regards,
Gun Ahn

I'm having trouble figuring out how to scale our data to input to this model.

The statement: " All signal are represented as 32 bits floating point numbers at the scale 1e-4V: so if the signal is in V it should be multiplied by 1000 before feeding it to the neural network model." Just doesn't make any sense to me. Does "at scale 1e-4" mean your A/D converter has a resolution (LSB) at the electrode of 0.1 milli-Volt (100 uV)? What is its bit depth?

We have ECG recordings in units of milli-Volts, what do I multiply by before inputting into your model? Our recordings are done at 1000 Hz and are 10 seconds long, resampling to 400 Hz and zero-padding to 4096 samples is not an issue.

Reading some tracings from your exams_part0.hdf5 database, there seems to be a large amount of baseline offset and "wander" within the traces. For example record ID 214626 (hdf5 index 1) Lead I starts at about 4.6 and drifts downward to about 3.2 whereas the maximum peak-to peak signal amplitude is about 1.4, are these units milli-Volts?

Getting an error for plot_learning_curves.py because train.py does not seem to append MAE to the dataframe. I am not able to tell where MAE is being calculated.

Will need to update the code, or include a max version requirement in requirements.txt for pandas<2*.