Whether we talk about EEG, ECG or EKG aren't all these signals discrete as we are reading the voltages from the electrodes that are due to certain bio-chemical or bio-electrical activities. I do not get why people say that EEG is continuous and hence we need to convert it to digital when it is already digital.

  • 1
    $\begingroup$ I don't understand why "bioelectric activities" can't take on a continuous range of values. The membrane potential of a neuron could be -70.123456789... That seems like it would be continuous to me. $\endgroup$ – Chuck Sherrington Aug 1 '19 at 22:00
  • $\begingroup$ So can a continuous signal be considered as a discrete signal with number of sample points tending to infinity? $\endgroup$ – Roshan Aug 2 '19 at 5:00
  • $\begingroup$ A continuous signal has a (countably) infinite number of points in it by nature. It's more that a discrete signal can be one with a finite number of samples of a continuous signal. When a cell depolarizes, it doesn't go from -70.002 to -70.001 by steps of 0.0001 or something, the voltage changes continuously. $\endgroup$ – Chuck Sherrington Aug 2 '19 at 21:14
  • $\begingroup$ @ChuckSherrington at the limit it changes by the total charge of all the ions that pass through the membrane in a unit time. $\endgroup$ – StrongBad Aug 3 '19 at 19:30
  • $\begingroup$ @StrongBad Not exactly. Ions can also be in different locations relative to the membrane. Unless space is also discrete it is truly continuous. $\endgroup$ – Bryan Krause Aug 3 '19 at 23:21

Biological signals are analogues and hence continuous. Early EEG systems simply recorded the analogue signals and displayed it as a continuous signal graphically as wiggly lines written by little pens on a roll of paper (Fig. 1). Only after analogue-to-digital converters became available, could EEG signals and the likes be digitally sampled.

analogue EEG signal
Fig. 1. Analogue EEG signal. source: Britton et al., 2016)

- Britton et al., Electroencephalography (EEG): An Introductory Text and Atlas of Normal and Abnormal Findings in Adults, Children, and Infants, 2016. Chicago: American Epilepsy Society

| improve this answer | |
  • $\begingroup$ i understand that it looks continuous but logically if you see that each point is a potential -> Discrete. $\endgroup$ – Roshan Jul 30 '19 at 7:48
  • 2
    $\begingroup$ Each point is a potential yes, but why would it be discrete? The original signal has no points, there are only discrete data points if you create them. $\endgroup$ – AliceD Jul 30 '19 at 8:15
  • $\begingroup$ @AliceD the potential comes from ions flowing through the membrane. These ions each have a discrete charge. $\endgroup$ – StrongBad Aug 3 '19 at 19:32
  • 1
    $\begingroup$ @StrongBad yes, but for a potential in the order of microvolts to be recorded in your average EEG signal, say, millions of electrons flow. With Coulomb's law you can calculate how much. Gross potentials recorded with superficial electrodes measure thousands of neurons, and even within a single neuron your supposedly fractal ion flow is anything but when recorded with an electrode. $\endgroup$ – AliceD Aug 4 '19 at 21:31

To add to the answer provided by @AliceD, pure digital waveforms are square waveforms as they represent steps between 1s and 0s and are therefore not continuous.

Digital waveform

Analogue waveforms are not. They are smooth continuous waves and can represent many voltage points at each millisecond, microsecond or nanosecond between the peaks and troughs.

analogue waveform

Outputs from ADCs (Analogue to Digital Converters) will have steps in them representing the various voltage ranges at their relative timings. In a sense they are digital outputs but are really a digitized version (a representation) of an analogue waveform through digital sampling of the wave voltage at the relevant millisecond, microsecond or nanosecond depending on the ADC timing (see the image below from https://www.allaboutcircuits.com/textbook/digital/chpt-13/practical-considerations-adc-circuits/).

digitized version of an analogue signal

| improve this answer | |
  • $\begingroup$ What's the point of having a digital ouput? Why can't we just take the sample times/points on the signal and use them for filtering/other operations? $\endgroup$ – Roshan Aug 2 '19 at 5:01
  • $\begingroup$ @Roshan Computers are digital. Hardware can do filtering before digitization and typically does when recording these sorts of signals to make them appropriate to digitize. $\endgroup$ – Bryan Krause Aug 2 '19 at 15:06
  • $\begingroup$ I am not a neurologist so therefore I don’t interact with EEGs etc. but as Bryan points out, computers are digital and cannot deal with analogue signals. If VGA/HDMI/DVI monitors are used, the signal will need digitizing and processed in order for the monitor to display it. $\endgroup$ – Chris Rogers Aug 2 '19 at 16:16

I disagree with the other answers, to an extent. Biological signals are generally created by ion passing through a cell membrane. Each of these ions cause a tiny current/magnetic field which is too small to measure in most cases. Each cell action potential is made up of hundreds (???) of ions passing through the membrane at a variety of positions. Most measures of biological signals measure potentials from many cells. These cells are distributed in a non homogeneous extracellular matrix. At the limit the biological signals are discrete and made up of individual ions, but by the time we measure them, they are so mixed up that they are continuous.

| improve this answer | |
  • 1
    $\begingroup$ Like I commented to your other comment, charged ions do not exist in a binary inside/outside configuration, they can also be partly between. Unless space is discrete, voltage is continuous. In any case, digitization is making a signal far more discrete than originally measured. $\endgroup$ – Bryan Krause Aug 3 '19 at 23:25
  • $\begingroup$ Do you have any references to back your claims @StrongBad? $\endgroup$ – Chris Rogers Aug 4 '19 at 10:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.