diff --git a/demos/matlab/See_octave_folder_for_matlab_demos b/demos/matlab/See_octave_folder_for_matlab_demos
new file mode 100755
index 0000000..600024f
--- /dev/null
+++ b/demos/matlab/See_octave_folder_for_matlab_demos
@@ -0,0 +1,2 @@
+See the octave folder for matlab demos (Octave is an open-source clone of Matlab).
+
diff --git a/demos/octave/SchreiberTransferEntropyExamples/runHeartBreathRateKraskov.m b/demos/octave/SchreiberTransferEntropyExamples/runHeartBreathRateKraskov.m
index d2fe09f..ddf27ce 100755
--- a/demos/octave/SchreiberTransferEntropyExamples/runHeartBreathRateKraskov.m
+++ b/demos/octave/SchreiberTransferEntropyExamples/runHeartBreathRateKraskov.m
@@ -6,7 +6,7 @@
 % 22/1/2014
 %
 % Used to explore information transfer in the heart rate / breath rate example of Schreiber --
-%  but estimates TE using Kraskov-Grassberger estimation.
+%  but estimates TE using Kraskov-Stoegbauer-Grassberger estimation.
 %
 %
 % Inputs
@@ -42,7 +42,7 @@ function [teHeartToBreath, teBreathToHeart] = runHeartBreathRateKraskov(kHistory
 	
 	fprintf('TE for heart rate <-> breath rate for Kraskov estimation with %d samples:\n', timeSteps);
 
-	% Using a single conditional mutual information calculator is the least biased way to run this:
+	% Using a KSG estimator for TE is the least biased way to run this:
 	teCalc=javaObject('infodynamics.measures.continuous.kraskov.TransferEntropyCalculatorKraskov');
 
 	for knnIndex = 1:length(knns)
@@ -52,27 +52,31 @@ function [teHeartToBreath, teBreathToHeart] = runHeartBreathRateKraskov(kHistory
 		% Perform calculation for heart -> breath (lag 1)
 		teCalc.initialise(kHistory,1,lHistory,1,1);
 		teCalc.setProperty('k', sprintf('%d',knn));
-		teCalc.setObservations(octaveToJavaDoubleMatrix(heart), ...
-					octaveToJavaDoubleMatrix(chestVol));
+		teCalc.setObservations(octaveToJavaDoubleArray(heart), ...
+					octaveToJavaDoubleArray(chestVol));
 		teHeartToBreath(knnIndex) = teCalc.computeAverageLocalOfObservations();
 		
 		% Perform calculation for breath -> heart (lag 1)
 		teCalc.initialise(kHistory,1,lHistory,1,1);
 		teCalc.setProperty('k', sprintf('%d',knn));
-		teCalc.setObservations(octaveToJavaDoubleMatrix(chestVol), ...
-					octaveToJavaDoubleMatrix(heart));
+		teCalc.setObservations(octaveToJavaDoubleArray(chestVol), ...
+					octaveToJavaDoubleArray(heart));
 		teBreathToHeart(knnIndex) = teCalc.computeAverageLocalOfObservations();
 		
-		fprintf('TE(k=%d): heart->breath = %.3f, breath->heart = %.3f\n', knn, teHeartToBreath(knnIndex), teBreathToHeart(knnIndex));
+		fprintf('TE(k=%d): heart->breath = %.3f, breath->heart = %.3f nats\n', knn, teHeartToBreath(knnIndex), teBreathToHeart(knnIndex));
 	end
 		
 	tElapsed = toc(starttime);
 	fprintf('Total runtime was %.1f sec\n', tElapsed);
 	
-	plot(knns, teHeartToBreath, 'rx');
+	plot(knns, teHeartToBreath, 'rx', 'markersize', 10);
 	hold on;
-	plot(knns, teBreathToHeart, 'mo');
+	plot(knns, teBreathToHeart, 'mo', 'markersize', 10);
 	hold off;
 	legend(['TE(heart->breath)'; 'TE(breath->heart)']);
+	set (gca,'fontsize',26);
+	xlabel('K nearest neighbours', 'FontSize', 44, 'FontWeight', 'bold');
+	ylabel('TE', 'FontSize', 44, 'FontWeight', 'bold');
+	print('results.eps', '-deps', '-color');
 end