Identifying setup components

Check if everything is there...

AND.... Check the connections

AND.... Got it all?

Start EEG_Recorder in MATLAB (32bits)

1. Check if WIFI is available, on the recording Laptop.  This is necessary for MATLAB license verification.
2. Start MATLAB (32bits)
3. Make the content of the EEG_recorder folder, the current folder. See example on the right.

Switch your Mobilab amplifier on

• This to be able to test your connections later
• You can find your Mobilab amplifier in the picture above

START THE PROGRAM

• - by typing ‘eeg_recorder’ in the command window
• - by dragging eeg_recorder.m to the command window
• - Via the edit Mode. DANGER: ONLY USE THIS OPTION IF YOU REALLY WANT TO EDIT. Double clicking eeg_recorder.m and then clicking on the green ‘Run’‐button. This opens the eeg_recorder.m file in the editor.

EEG_recorder contains the following main functions:

• Record EEG
• Generate a live view of EEG in time domain and frequency domain
• Saves/Loads EEG to/from RAM and hard disk
• Analyze EEG data using several toolboxes

CONNECTING the ELECTRODES

Read through the connecting procedure below, BEFORE actually starting it

1. Select a subject to record from
2. Find the GAMMABox as part of the setup. Connected are:
• 4 red recording electrodes
• 1 yellow ground electrode.
• 1 blue reference electrode.
3. Sticker the 4 recording electrodes
4. Place two recording electrodes on the right arm:
• Stick electrode 1 on the Extensor Digitorum (image 7) Find it by making a fist and move your wrist upwards.
• Stick electrode 2 on the Palmaris Longus (image 7). Find it by making a fist and move your wrist downwards.
5. Connect electrode 3 on the left wrist. Turn your palm up, so your thumb points to the left. Stick the electrode about 3 cm beneath the wrist joint, left of the central tendons, the location where you can feel yoiur heart beat.
6. Now connect electrode 4at a central location on the fore head, with the wire pointing horizontally to the right.
7. Stick the yellow ground elctrodenext to the central PFc electrode on the forehead, with the wire pointing horizontally to the right.
8. Tape al the wires to tye skin, releaving the tension on the electrode.
9. Stick the wires behind the right ear of the test subject.
10. Fill tye syringe with the contact gel and fill the electrodes with the gel (see image on the left).
11. Place a small drop of gel on the blue reference electrode and clip it on the right earlope (see image on the right).

Image: Gelling electrodes.

Image 7: The muscles of the arm. The activity of the Extensor Digitorum and the Palmaris Longus are recorded in this experiment.

SETTING the RECORDING-COMPUTER Module II :

When electrodes are connected, the signal can be tested. When looking at the recording program, all EEG and all marker channels are switched on by default. "Just to be save".

Swich of the ones that we dont need now. (see image 8).

Image 8: Selecting EEG and marker channels in EEG_Recorder.

1. Activate channels 1 through 4, turn off channels 5 through 8 . See Image! 4 red electrodes are connected so 4 channels need to be recorded. 
2. NOTE: the yellow ground and the blue reference electrodes just serve as ground and reference for all recording elctrodes.
3. Turn all markers (DIOs) off, See image 8. For now we dont use them.

Image 9: The interactive ‘Live view’ panel of the EEG_Recorder.

1. ‘Preview sample length’ = 1024 samples. So, 1024 data points are shown on the screen. With a sampling rate of 256 samples/s, this allows 4 seconds of data to be pushed to the screen.
2. ‘Channel spacer’ default = 1. The smaller the channel spacer, the larger the amplitude of te signals. Keep the Channel Spacer to 1.
3. Keep the ‘FFT sample length’ equal to the ‘Preview sample length’. The FFT is than calculated over the signal part that is shown on the screen.
4. The ‘FFT frequency range’ is set from 0.1 to 60 Hz

Image 10: The Clear/Start/Stop menu of the EEG_Recorder.

The ‘Time to record (s)’ marks the predetermined recording time.

1. Default = 3600 seconds, 1 hour before the recording stops automatically.
2. Then acquisition stops and you can save the data.
3. However, you can also stop the recording on your own time. (see below: Stop EEG).
4. Note: sample frequency = 256 samples/s 

• CLEAR: Before starting a new recording CLEAR clears the buffer (your temporary datastorage) and the screen. Your settings remain.
• BEWARE: Never use CLEAR when you havent saved the data: You loose everything.
• Start EEG: Acquisition starts.
  • BEWARE: Be patient. It takes a while (seconds) before the signal actually appears on the screen and the green LED on the MOBILab light up continuously.
  • The recording stops automatically when the ’Duration acquisition’-time is reached.
  • The entire recording of all 8 channels (also the ones that are not connected) are presented on the screen
  • SAVE the data: File > Save.
  • BEWARE: Not saved is all data gone! So dont hesitate: SAVE the data. After that you have time to discuss, talk, get coffee etc etc.
• Stop EEG: When you want to stop the recording at your own free will, before the ‘Duration acquisition’ is reached.
  • BEWARE: Be patient. It takes a while (seconds) before the recording actually stops. Only when the entire recording of all 8 channels appears on the screen, it is ready to be saved. Wait for it before doing anything else.
  • SAVE de data: File - Save, in the folder that is automatically opened, the data folder.
• Save to Ram & ... is the default and does not need changing.

First have a look at both EMG signals

1. Start a new recording
2. Evaluate the signaal. Still OKAY?
3. Ask the test subject to move the wrist up and down, in alternation.
4. Observe the signal and discuss the results.
5. Now ask the test subject to make a fist with maximum force.
6. Observe the signal and discuss the results.

Now have a look at the ECG signal

1. Start a new recording
2. Evaluate the signal: Still Okay, no 50 Hz?
3. The ECG signal is best when the left arm rests on the table.

Nuow its time to focus on the PFC electrode

1. Start a new recording
2. Evaluate the signal.
3. Ask the test subject to sit up and look ahead.
4. Observe the signal
5. Ask the test person to blink 5 times in a row.
6. Observe the results and discuss.
7. You recorded a EOG, an electro oculo gram.

You have recorded EMGs ECGs as well as EEG. In Module III you are going to use this in a actual sensori-motor experiment

Dit experiment wordt uitgevoerd door één deelnemer van de groep

Extra materiaal:

• 1 tennisbal of gebruik de geltube om in te knijpen

De electrodeconfiguratie is zoals tijdens het testen van de signalen.

Sluit een proefpersoon aan! 

Test nu eerst het signaal

• Start de EEG_recorder weer op (zoals eerder)
• Maak eerst het signaal zoveel mogelijk storingsvrij (trouble shoot als nodig).
• Het signaal moet er ongeveer uitzien als in afbeelding 11, met een extra spierkanaal.
• Als het signaal naar tevredenheid is stop dan de meting, CLEAR de buffer en het scherm.

Afbeelding 11: Zoals de meting eruit kan zien. Hier kanaal 1: hartslag, Kanaal 2: EMG en kanaal 3: Frontaal met Oogknippers

Doe nu het feitelijke experiment

Geef nu de proefpersoon de tennisbal of de geltube in de rechterhand. Laat de tube op de tafel staan en knijp de tube losjes tussen de wijsvinger en de duim. Hou de andere vingers in een vuist. Laat de proefpersoon nu naar een leeg/wit scherm naar de in het midden geplaatste cursor kijken om oogbewegingen zoveel mogelijk te voorkomen en verstoringen te vermijden.

• Start een nieuwe meting met een duur van 1 uur .
• Wacht eerst minimaal 60 seconden voor een stabiele ‘baseline’, zonder de tube al in te knijpen.
• Laat vervolgens de proefpersoon zo hard mogelijk in de tube knijpen en dit 5 minuten volhouden (dus houd de tijd bij).
• Niet even loslaten maar continue blijven knijpen. 
• Ontspan na 5 minuten, of eerder als de proefpersoon niet meer kan, maar hou de tube nog wel lichtjes vast.
• Meet nadat de proefpersoon gestopt is met samenknijpen nog minimaal één minuut door.
•  
• Daarna wederom knijpen maar nu elke minuut even loslaten en meteen weer knijpen. 
• Laat de proefpersoon dat 5 keer herhalen
• Meet uiteindelijk nog minimaal één minuut door.  
• Stop de meting.
• SAVE de data!
• Overleg even met je assistent of er nog tijd is om dit experiment te herhalen met andere personen in de groep.
•  
• Deze data gaan we analyseren in de eerste laptop werkgroep.
• Zorg daarom dat iedereen in de groep de data heeft!

The experiment is performed by one member of the group. Preferably some else than in the previous recording

Now we are going to use the full setup that is place in front of you; Materials:

• Two laptops (Recording and stimulation laptop)
• Five recording elektrodes, and one ground and one reference electrode
• Recording equipment
• EEG-gel, elektrode stickers, tape

STIMULUS COMPUTER

To record the key-presses and couple them to the muscle potentials we use a Matlab Go-NoGo program that has been made especially for this practical 'GoNoGo.m' that is present on the stimulus computer

The Experimental setup and what the task sends to the EEG_recorder

• The onset of each and every tone is detected and passsed to the EEG recorder
• The onset and identity of every keypress is detected and passsed to the EEG recorder

Logging the events

• All events of your choice are transferred to digital pulses
• These pulses are send to the Digital Input/Output (DIO) box (National Instruments: USB-6501)
• The marker (TTL) is send to the EEG-recording laptop via a predetermined (by the task developer) gate on the USB-6501.
  • TTL, short for Transistor-Transistor Logic is a standard digital pulse for communication between electronic equipment and circuits
  • The receiver reads the signaal (0= maximum 0.8 Volt, 1 = minimal 2.0 Volt).
• In this experiment, the markers indicate the onset of the keypresses and tones as recoded by the task laptop.
• The analysis software, as part of the EEG_recorder program, can cut the EEG according to the markers and produce ERPs.
• The individual ERPs can be averaged into a proper ERP related to a particular event.

Of course the ERPs only make sense if controlled by a proper task that delivers meaningful stimuli.

For now, the GoNoGo experiment is up and running and preprogrammed. However, you have to design your own task that suits your own experiment.

• Download the GoNoGo folder from CANVAS and unzip on the desktop.
• Start MATLAB 64-bits.
• Make The GoNoGo folder your Current Directory in MATLAB
• Double click the GoNoGo.m file, it opens in the editor.
• Now first read through the code and notice what it does.

RECEIVING MARKERS IN YOUR EEG-RECORDING

Each time a key is pressed and a tone is played on the task laptop, TTL pulses are send to the EEG recording laptop via the USB-6501. Only when the EEG recording laptop is set to receive, the markers will actually be stored with the EEG recording (Picture 13).

Picture 13: TTL-puls panel in EEG_recorder. Select the markers that have beeen set in the task laptop. In the sensori-motor experiment you need all six                     .

Test the connection of the task laptop with the recording laptop

Before you perform the experiment it is essential that you test whether everything is set up properly. You van do this by performing a marker test

Selecteer op de stimulus laptop de Marker test!.

• Start MATLAB 64-bits on the task laptop.
• Open the script Test_Markers_NI_USB_6501.m in the editor (double click).
• Have a look at the script and try to get an idea what it is going to do
• Start the EEG-Recorder without connecting a test subject
• Now run the script.
• The command line of MATLAB shows the progress.
• The connection with the USB6501 is configured and the marker channels are loaded.
• Watch the EEG-recorder recording and verify that it is doing what you already figured out.

The Marker test program stops automatically.

STOP the EEG_Recorder and SAVE the file    

Now observe the saved file, that is still visible in your recording window.

Verify that the markers are actually recorded by the EEG_recorder.

In case of a succesfull marker test, the real sensori-motor experiment can start

Make The GoNoGo folder your Current Directory in MATLAB

Double click the GoNoGo.m file, it opens in the editor.

Get the EEG-Cap, call one of the teachers and have them explain how to place the electrodes

Choose a new test subject.

• Three red EEG electrodes are connected
• Place at locations FZ, C3 and PZ
• Yellow ground on CZ
• Place the EEG cap on the head of the subject.
• Measure Inion and Nasion (see image 14) and fit the cap in line with the international standard, the 10-20 system.

Image 14: Left the location of CZ elektrode with reference to Nasion and Inion. On the right the standardized placements of electrodes.

• Place electrode 4 on the Digitorum extensor
• Place electrode 5 op the Palmaris longus
• Fix the wires with some tape to release the pressure on the electrode.
• Clip the reference electrode on the earlobe.
• The yellow ground is now placed in the Cap on Cz
• Gel all electrodes (See Image 15)

Afbeelding 15: Het ‘gellen’ van de elektrodes.

• Run the EEG_Recorder while stabilizing your signal.
• A proper ERP depends on the quality of your signal.

Prepare the EEG-Recorder

• One of your team is responsible for the recording during the experiment
• Switch ON EEG channel 1 through 5
• Switch On all six marker channels
• Start The EEG_recorder.
• Check the signal.
• Adapt if neccesary.
• Stop, clear and start the EEG_recorder again

Start the Go-NoGo task on the Task Laptop.

• One of your team is responsible for everything concerning your test subject during the entire experiment
• Run the Go-NoGo task
• The connection with the USB-6501 is made and the marker channels initiated.
• The experiment starts running

The experiment runs for about 20 minutes.

• Stop your EEG recording
• BE PATIENT: Wait for the download of the buffer to the screen, this can take sime time after a long experiment
• NOW, SAVE YOUR DATA, immediately. Save first, discuss later.
• KEEP AWAY from the CLEAR BUTTON AF!.

Release your test subject

Copy all the recording files to every member of the team.

NOTE: You can't email .mat files. These are considered virus susceptable. You need to zip the files before distributing.   

• The recording file is located on the EEG-Recorder Laptop
• In the folder where you saved the file.
• Most probably in the data folder of the EEG_Recorder program on the desktop
• De behavioral file with the identity of the key presses etc etc is on your task laptop

ANALYSIS TIME

Here we focus on the analysis of the EMG signals. The analysis of the EEG recordings follows similar steps, however, not all parameters are the same so think before you do!

Als je onderstaande stappen hebt doorlopen en begrepen, verwijzen we voor de analyse van het EEG in relatie tot de tonen naar de ERP_ANALYSIS pagina. Daar staat meer uitleg over het hoe en waarom

Loop nu onder de stappen door om het EMG te analyseren.

DATA ANALYSIS in 6 logical steps

Data analysis usually is performed in a few logical steps.

• Filter out unwanted frequencies in your data
• Cutting the data on the basis of the marker onsets: create individual ERPs
• Baseline correction to 'zero' the baseline of individual ERPs
• Select relevant ERPs
• Average the cut and corrected ERPs
• Visualization of the averaged ERPs

Below each step is dealt with in order.

STEP 2: CUTTING THE DATA 

The second step is to cut the data in short segments on the basis of the onset of the keypresses. The tool for this is the ‘Event cutter’.

• Open ‘Event cutter’ in the EEG_recorder tools. It looks like the image below


Image : Event cutter Tool. Used to cut EEG data in ERP segments on the basis of marker events.

Each time a key is pressed a TTL-pulse is send to the recording laptop (the DIO's or markers). We would like to cut the data based on these TTL-marker pulses. Therefor you have to load the data first, the filtered data file that you have just created.

• Load the data: File -- load.

We are using the 'TTL Based cuts' section. The TTL-pulses of the keyboard and the onset of the tones are saved in channels 9 - 14. The first eight are always your EEG channels and then one extra channel for each marker channel. So you have a maximum of 14 channels (8 EEG/EMG channels and 6 marker channels).

If the sensori-motor experiment is properly executed, you actually have data on all 6 marker channels. Have a look at the explanation of the Go-NoGo task what is actually recorded.

• Select channel 14, in the experiment this is the channel onto which the keypresses of condition 2 are recorded. In Matlab the data of the individual channels are stored in columns. So file in 14 as column number (see image 16).

We are now first interested in the reaction of the muscles in relation to the keypresses.

• Therefore, Cut the data 256 samples (1000 ms) before the marker onset and 128 samples (500 ms) after the marker onset. NOTE: The unit here is ‘samples’ (What was the sampling frequency of the EEG system?).

• Press ‘cut’, Now you cut short pieces of data from each EEG channel surrounding the onset of the keypress on the basis of the set parameters.
• Save the data and use the original name with the extension _cut (Never use spaces or signs other than letters, numbers and underscores, Matlab is allergic to that). Save into the datafolder on your desktop.

STEP 3: BASELINE CORRECTION OF THE ERPs

Now close 'Event_cutter’.

The next step is to correct the baseline for each and every ERP that you have produced. You can use the ERP tool for that (Indeed, you can find ir in the toolbox).

• Open ‘ERP tool’ in the EEG_recorder tools.


Image : ERP_Tool. For adjusting and analysis of ERPs.

• Load the file that you just saved and which houses the ERPs (File -- Load).
• In the green section: The first number represents the amount of samples of each ERP, the second number is the amount of channels. In the (new) third dimension and all the individual ERPs that were just cut. We would like to adjust the baseline for all these ERP sections.
• In the blue section
• Channels to browse: leave that on '1'. The baseline correction is automatically performed on all channels.
• Onset (in sample nr!): provide the length of the baseline. Basically the amount of data before the marker onset.
• Fill in : 256 samples. This is one second with the sample frequency of 256 samples per second
• ONSET in sample nr: 256 samples. This is one second with our sample frequency of 256 samples per second. We hereby take the first second of the each ERP as the baseline or control measurement.
• The principle of baseline correction: We took many 1.5 seconds sections of data from the file. If you consider the fact that there can be a low frequency oscillation with an amplitude of 100 microvolt and a frequenty of 2 Hz runs through your data than some ERPs are cut when the low oscillation is at its peak and some when the low oscillation is in a trouph. The first case the ERP starts at 50 microVolt, in the second at -50 microVolt. The baseline correction takes the average of the amplitude of the first second of each ERP, determines the deviation from the '0' line and corrects the amplitude of the entire ERP with that number.
• Click on Baseline Correction: Each ERP, for every channel is now corrected with there own baseline values.
• Save the data Use the old name with the extention _bl. Save to the datafolder on your desktop.

The data is now baseline corrected.

STAP 4: SELECTEREN VAN DE JUISTE MARKERS

De volgende stap is het selecteren van de juiste events binnen de geknipte stukjes EEG data.

NOTE: Na alle tonen werd er ongeacht of er een key wordt ingedrukt een 'keypress' marker gecreëerd. Eigenlijk geeft de marker het 'end of trial' aan. Dus de ERPs die geknipt zijn zijn een mix van werkelijke keypresses EN het aflopen van een trial zonder dat er een keypress heeft plaatsgevonden. Haal uit je gedragsfile (de .csv file) welke markers werkelijk gerelateerd zijn aan een keypress. Deze ga je nu selecteren door de 'end of trial' stukjes data te verwijderen.

• Open ‘Array Manipulator’ in de EEG_recorder onder tools. We gaan nu rechtsonder 'Selection' tool daarvan gebruiken


Afbeelding : De selection tool binnen de Array_manipulator_Tool. Voor het selecteren van stukken data uit je file.

• Laad eerst de zojuist gemaakte basislijn gecorrigeerde file (File -- Load).

De eerste twee dimensies van je dataset zijn meetpunten in de tijd en de verschillende kanalen. In de (nieuwe) derde dimensie zitten de verschillende stukjes die we net hebben geknipt. Hier gaan we nu alleen de werkelijke keypresses selecteren door de 'end of trial' markers te verwijderen.

• Kies daarom ‘Discard’ en ‘3rd dimension’
• Vul nu alle trial nummers in die NIET geleidt hebben tot een keypress, gescheiden door een comma. Dus bv. 1,3,5,18,24
• Click op 'Apply Selection'
• De derde dimensie, links bovenin de 'array manipulator' tool, geeft nu aan hoeveel trials er zijn overgebleven!
• Sla deze data op onder de oude naam_filt_cut_bl met de toevoeging _keypresses (gebruik geen spaties, Matlab kan daar niet tegen). Sla het op in je datafolder op de desktop.

De data bestaat nu alleen uit geselecteerde trails met keypresses.

STEP 5: AVERAGING THE DATA

The next step is averaging the data. For that we are going to use the Array_manipulator (found in the tool box).

• Open ‘Array Manipulator’ in the EEG_recorder tools. We are using the 'Average along dimension' tool


Image : The 'Average along dimension' tool as part of the Array_manipulator_Tool. Use to average data within a file.

• Load the file with the selected keypresses (File -- Load).

The first two dimensions (columns) of the dataset are the data points in time and the different channels. The third dimension hold the various ERPs that have just been baselined. These will now be averaged.

• Choose ‘Average along dimension:’ ‘3rd dimension’ and press ‘Average’.
• The third dimension disappears (is now 1 average).
• The averaging has taken place for all 8 channels. NOTE: So your channels of interest as well as the rest of the channels are averaged.
• Save the data Use the old filename name_filt_cut_bl and add _keypresses_avg (SAgain, no spaces, Matlab doesn't work well with spaces). DSave in the data folder on your desktop.

The data is now averaged over all ERPs.

STEP 6: PLOTTING THE DATA

The result of it all is 8 channels with 1500 ms of dta each: a 1000 ms before and 500 ms after the onset of the keypress.

• Now close ‘array manipulator’.

Now it is finally time to look at the result. Use Data_plotter in the toolbox.

• Open ‘Data plotter’ in the Tools menu


Image : The Plotting tool. To plot your data.

• Load your data_cut_bl_keypresses_avg file.
• Select 'Plot Columns' to plot one average ERPs. Remember that Matlab has stored each channel data in its respective column.
• Set The ‘Onset (s)’ on 1 s (This sets the '0' of the ‘X-axis’ on the time in the ERP that holds the onset of the keypress.
• Note: this 'Onset' setting is completely dependent on the pre marker settings that you chose during the cutting process!!
• Select in 'Rows/Columns to plot', both channels that hold your EMG signals!
• Press ‘Plot’.
• Maximize de plot
• AND?!.......

This ends the SENSORI-MOTOR EXPERIMENT: Thank you and good-bye!