Label predicted seizures in EEG data

I used a deep learning model to predict seizures in EEG data. I want to plot a graph to show where the model predicted the seizures on the EEG data and where the actual seizures are. The predicted values are categorical values 1 for seizure and -1 for no seizure. The EEG signal can be ploted using plot(channel).

Answer 1

[Image Analyst]

You could shade the zones where a seizure was detected, or you could just plot the seizure in a plot below your data.

subplot(2, 1, 1);
plot(eegData, 'b-', 'LineWidth', 2);
title('EEG Signal')
grid on;
subplot(2, 1, 2);
plot(predictedValues, 'b-', 'LineWidth', 2);
grid on;
title('Prediction')

[LOMenz]

Thank you. I would like to shade the predictions on my EEG signal. How can I shade please?

Answer 2

[Image Analyst]

You can use patch():

numSamples = 30;
x = linspace(0, 2*pi, numSamples);
eegData = 5 + 10 * sin(2 * pi * x / 3);
threshold = 6;
predictedValues = eegData > threshold;
subplot(2, 1, 1);
plot(x, eegData, 'b.-', 'LineWidth', 2, 'MarkerSize', 18);
yline(threshold, 'LineWidth', 2)
yl = ylim;
title('EEG Signal.  Red Zones = Seizures')
grid on;
subplot(2, 1, 2);
plot(x, predictedValues, 'b.-', 'LineWidth', 2, 'MarkerSize', 18);
grid on;
title('Is Seizure')
ylim([0, 1.2])
% Shade regions
subplot(2, 1, 1); % Switch back to the EEG axes.
for k = 2 : numSamples
    if (predictedValues(k) == 1 && predictedValues(k-1) == 1)
        % Shade this region
        x1 = x(k-1);
        x2 = x(k);
        y1 = yl(1);
        y2 = yl(2);
        xData = [x1; x1; x2; x2; x1];
        yData = [y1; y2; y2; y1; y1];
        patch(xData, yData, 'r', 'FaceAlpha', 0.2, 'EdgeColor', 'none')
    end
end

[Image Analyst]

See attached demos.

[LOMenz]

Thank you, this was very useful.