![\"MIT](\"https://img.shields.io/badge/License-MIT-brightgreen.svg\"){kind=icon}
\n\n![](\"https://img.shields.io/github/v/release/kylemath/EEGEdu?include_prereleases\")
\n\n![\"PRs](\"https://img.shields.io/badge/PRs-welcome-brightgreen.svg\"){kind=icon}
\n\n![\"issues\"](\"https://img.shields.io/github/issues-raw/kylemath/EEGEdu\")
\n![](\"https://img.shields.io/github/issues-closed-raw/kylemath/EEGEdu\")
\n\n![](\"https://neurotechx.herokuapp.com/badge.svg\"){kind=icon}
\n\n\n![](\"https://img.shields.io/badge/NeurotechXChannel-%23eegedu-lightgrey\"){kind=icon}
\n\n \n\n\n\n\n\n\n\n\n
\n![\"Interactive](\"logo.png\")
\n
\n ![\"Project](\"screenshot.png\")
\n
{translations.footer}\n A \n Ky\n Kor\n Key \n Production.\n Github source code \n
\n{specificTranslations.description}
\n{\n \"Last module we recorded the frequency bands while people had their eyes open or closed for 10 seconds. Although we predicted there should be larger alpha power when the eyes are closed, are results showed the opposite effect:\"\n }
\n{\n \"We pooled our estimates for alpha in eyes open vs eyes closed as a class, and computed the average and compared the results with a paired samples t-test. As you can see, even if we remove some of the outliers with extreme values that may have been artifacts, we still find that for most people we estimated larger alpha power when their eyes were open. Why could this be?\"\n } \n You can access a read only copy of the spreadsheet we used to compute these group level effects here. \n
\n{\n \"We discussed in class that we had very little control over WHAT people were doing while their eyes were open. Some were looking around, some may have been blinking, some may have been reading their screen. There was a great deal of variability in the EEG signal due to differences in peoples behaviour during our task. As we have seen in previous modules, these kind of artifacts can create artificial power in most lower frequency bands. Because we were looking just at the alpha frequency band, we did not estimate the full range of differences across frequencies. \"\n }
\n{\n \"Furthermore, the short recording time (10 Seconds) also allows for less reliable measurement of the EEG signal. In addition, people may not have been prepared when the button was pressed to start their recording. \"\n }
\n{\n \"Based on all these ideas, in this module we are going to run a more controlled experiment comparing the EEG during eyes open and eyes closed conditions. Instead of just recording averages over frequency bands, we will record the whole spectra.\"\n }
\n{\n \"This module, the recording sessions will be 60 seconds long. They will also begin with a 2 second delay after pressing the button. A window will pop up with a red fixation cross. In the eyes open condition participants should keep their eyes fixed on this red fixation cross the entire 60 seconds, and should try to minimize blinking as much as possible.\"\n }
\n{\n \"Please DO NOT change any of the settings. The module will pick a random number to tell you which condition to run first, so that we can counterbalance the order. Keep track of this order in a lab notebook you will submit.\"\n }
\n{\n \"Each person will record two sessions, one with eyes open and one with eyes closed. The output files are identical to those we recorded in the spectra module 5. \"\n }
\n{\n \"The random number generator suggest you run the following condition first: \" + firstType\n }
\n{\n \"Instead of averaging over all electrodes, we will separately average the frontal and temporal electrodes to look for differences in different locations on the head. First you will each analyze your own data to summarize what you found. We will average the spectra values over the entire 60 seconds, and also over contralateral pairs of electrodes. You can watch a video showing you how to do this if you need a reminder here: \"\n }
\n{\n \"Finally, we will paste the resulting average spectra in each condition, at each of the two electrode locations into a group spreadsheet that we will use for our final assignment paper in the class. To clarify, each participant will make two recordings 1) Eyes open, and 2) Eyes closed. For each, the analysis will produce two arrays of numbers (spectra), one for Frontal electrodes, and one for posterior electrodes. Therefore there are four different tabs in the following group data log:\"\n }
\n\n Your data is currently recording, \n once complete it will be downloaded as a .csv file \n and can be opened with your favorite spreadsheet program. \n Close this window once the download completes.\n
\n\n Your data is currently recording, \n once complete it will be downloaded as a .csv file \n and can be opened with your favorite spreadsheet program. \n Close this window once the download completes.\n
\n{specificTranslations.description}
\n{[\n \"We are working on allowing for people to make their own animations with a p5.js editor. \", \n \"For now if you want us to make a new one make a github issue with your p5.js animation and describe your idea\"\n ]}
\n{[\n \"Press connect above to see the chart.\" \n ]} \n
\n{[\n \"In the next demo we look at the traditional frequency bands. \",\n \"Oscillations in the brain are important as a mechinasm of brain function and communication. \", \n \"For example, within a brain area, cellular firing becomes locked to the ongoing oscillations of the local field potential: \"\n ]}
\n{[\n \"Since oscillations can control the timing of neural firing, they can also be used to communciate information and create distributed representations. \",\n \"Two nearby brain regions that oscillate in sync will have cells that also fire in sync, which also means the neurons they connect to will be more influenced. \", \n \"Different brain regions have different frequency oscillations at different times \"\n ]}
\n{[\n \"Oscillations in the brain seem to belong to a number of basic families or frequency bands that are influenced by cognition in different ways. \",\n \"We take the same spectra that was computed in Spectra and divide into five bands. \",\n \"Delta (1-4 Hz), Theta (4-7 Hz), Alpha (7-12 Hz), Beta (12-30 Hz), and Gamma (30+ Hz). \"\n ]}
\n{[\n \"These different frequency bands are associated with different brain states. \", \n \"For example, when we pay focus attention to something, our alpha goes down and our beta oscillations increase. \",\n \"Gamma oscillations are associated with neural activity, but for the most part Gamma oscillations are very difficult to measure outside the head (very small) and so we will ignore them. \",\n \"Theta oscillations increase during spatial navigation and are associated with learning and memory. \",\n \"Alpha oscillations are assocaited with attention and active inhbition of neural activity, we will consider them further below. \"\n ]}
\n{[\n \"Connect a muse and watch the following bar chart of the frequency band power. There is bar for each electrode. \",\n \"See if you can pick one of the frequency bands and try to control the height by relaxing.\"\n ]}
\n{[\n \"One of the earliest and easiest to measure changes in the EEG is that of alpha oscillations when the eyes closed. \",\n \"We will test these changes by recording data in two conditions, and comparing the average alpha at all four electrodes between conditions. \", \n \"We expect to replicate the following relationship: \"\n ]}
\n{[\n \"First go to the Raw module 3 and check the data and connection quality. \",\n \"Then come back to Module 6, no need to change any settings. \",\n \"We will record two sessions for each person in your group, one with eyes open and one with eyes closed. \",\n \"Once recorded you can open the .csv file and observe what gets saved. Along the rows are the different frequency bands from each electrode, we are going to average over all four electrodes in this assignment. \",\n \"We are also going to average over time, which is shown on different rows. \", \n \"So please compute the average ALPHA power in that output file, and do the same for the other condition, make sure to keep track of which file was created during which condition. \", \n \"Compare your values for eyes open vs eyes closed, did you find the expected difference? \",\n \"Why do you think alpha differs when we close our eyes? \"\n ]}
\n{[\n \"Once you are complete, move on to the next Module and control live animations with the values of these frequency bands. \",\n ]}
\n\n Your data is currently recording, \n once complete it will be downloaded as a .csv file \n and can be opened with your favorite spreadsheet program. \n Close this window once the download completes.\n
\n{specificTranslations.description}
\n\n {\"Clicking this button will begin the experiment so check your data quality on the raw module first. \"}\n {\"A window will pop up when you click the button and a series of circles will appear. Stare at the cross in the center. \"}\n {\"There will be red and blue circles, ignore the blue ones.\"}\n {\"Whenever you see a red circle, as fast as you can either press spacebar on a keyboard, or tap the touchscreen on a tablet or phone. \"}\n {\"This entire time the eeg data will be saved along with a column indicating which target was on the screen, and another for the responses. \"}\n {\"The task will continue for a few minutes and once it is finished a .csv file will automatically download. \"}\n {\"This .csv file has a row for each time point, a column for each electrode, and the columns indicating when targets appeared, and when responses were made. \"}\n {\"It saves a marker of 20 when the target is on, a marker of 10 when the blue standards are on, and a peak when the spacebar or touchscreen are pressed, here you can see those synced with an eeg channel: \"}\n\n
\n\n \n
\n Your data is currently recording, \n once complete it will be downloaded as a .csv file \n and can be opened with your favorite spreadsheet program. \n Close this window once the download completes.\n
\n{[\n \"As a first introduction to measurement of electrical potentials from the body we will look at something accessible. \",\n \"As the heart beats and pumps blood throuhgouts our body, a series of electrical potentials are created, which can be measured using electrodes placed around the heart. \",\n \"This is referred to as the Electrocardiogram (ECG). You have seen this hundreds of times in movies and TV beside hospital beds. \"\n ]}
\n{[\n \"The ECG is best measured by comparing the electrical potential accross the left vs. right side of the body. \",\n \"Depending on where on the body the measurements are taken, they pick up a different view of the electrical potentials generated by the heart. \",\n \"One of the easiest places to measure is comparing the voltage between the left and right hands. \",\n \"For example, this is how treadmills can read your heart rate when you place one hand on each of the two holds. \"\n ]}
\n{[\n \"Therefore, you can see your own ECG today using the muse. \",\n \"You can take off the muse from your head and place a finger on your right hand on the muse's reference electrode (in the center of the forehead). \",\n \"We can then place a finger of our left hand on one of the eeg electrodes. Lets use the left forehead electrode (position AF7). \",\n \"So place your left fingers pinching the left forehead electrode, and your right fingers pinching the center electrode. \",\n \"Otherwise try to hold the Muse as still as possible, and relax your body. \" \n ]}
\n{[\n \"If you haven't already, connect your muse and try to plot your heart rate before moving on to the data recording and analysis below. \",\n \"Soon you should see spikes in voltage measured between those two electrodes each time your heart beats. \", \n \"Watch what happens when you move your body or fingers, and notice how delicate the signal is. \" \n ]}
\n{[\n \"Here time is on the horizontal axis, and the voltage is on the vertical axis. \",\n \"There are 10 Seconds of data shown, with the current time shown on the right. \",\n \"There is a vertical line every second so you can estimate your heart rate roughly. \",\n \"Below we will save data and use it to estimate your heart more accurately in a spreadsheet. \",\n \"the signal will be red if it is noisy, and when you relax and hold still it will turn green/black\"\n ]}
\n{[\n \"Clicking this button will immediately record a 10 second long segment of data just like it is shown in the plot above. \",\n \"Therefore, make sure the chart above looks clean and you can see you heart beat clearly before pressing record. \",\n \"We are going to compare two conditions that show a clear difference in heart rate due to blood pressure changes: \",\n \"Standing and Sitting. \",\n \"You will record two sessions, one standing and one sitting, pick the order randomly but keep track of which output file is which\"\n ]}
\n{[\n \"A .csv file will be saved that can be opened in your favorite spreadsheet software like Microsoft Excel or in our examples, Google Sheets. \",\n \"Remember for each person at your computer to record two files, one while they are standing and one while they are sitting. \", \n \"Here is an example of what the data will look like once loaded. \",\n \"The top row shows the millsecond ellapsed in the data segment from 4 to 10,000. \",\n \"The bottom row shows the voltage measured from the difference between the two electrode, for each of those moments in time. \",\n \"The plot is plotting the data for each time point, this time with zero on the left. \" \n\n ]}
\n{[\n \"The following youtube video will show you how to open the file in Google Sheets, rename it, plot the data, find the peaks, \",\n \"Record them and find their difference in time, then take the average difference to estimate your average heart period. \",\n \"This value is then used to estimate your heart rate in beats per minute. \"\n ]} \n \n Link to example google sheet from video. \n \n
\n{[\n \"Finally each of you will enter this number for both sitting and standing into a google sheet that we are sharing as a class. \", \n \"We will use this shared google sheet which combines all our data in order to compute group statistics. \" \n ]}
\n\n Your data is currently recording, \n once complete it will be downloaded as a .csv file \n and can be opened with your favorite spreadsheet program. \n Close this window once the download completes.\n
\n{[\n \"Press connect above to see the chart.\" \n ]} \n
\n{[\n \"In the previous module we each estimated our heart rate in two conditions, while we were sitting, and while we were standing. \", \n \"We used a shared google sheet which combines all our data in order to compute group statistics. \",\n \"The following video shows how to use the data to make plots of the data, compute statistics, and test the difference. \" \n ]} \n \n A copy of the anonymized data that you can use to follow along with the video can be found here. \n \n
\n{[\n \"There are a few problems with the way we estimated heart rate in the previous module. \",\n \"We used only a single 10 second segment of data which adds variability. \",\n \"We also used peak detection to find each heart beat, which takes alot of time and is difficult with noisy data. \", \n \"This required us to make arbirary thesholds to find peaks, and led to some poor estimates in heart rate, as indicated by the number of outliers we needed to remove (~15). \", \n \"Therefore in this module we will use some signal processing to get a better estimate of heart rate. \" \n ]}
\n{[\n \"Here is an example of some ECG data from our experiment in Module 2. \",\n \"Notice that just like many other aspects of our bodies function, our heart rate is a rhythm, that is, it repeats in time at regular intervals. \"\n ]}
\n{[\n \"Therefore, we can use mathematical techniques such as a fourier transform to estimate what frequency is present in the ECG data. \",\n \"A fourier transform turns any series of numbers into a summed set of sine waves of different sizes. \",\n \"The following animation shows how a single time-series of data, can be thought of as the sum of different frequencies of sine waves, each of a different magnitude. \", \n \"The blue line chart in the animation shows what is called the spectra, and indicates the power at each frequency.\" \n ]}
\n{[\n \"Now we can use the muse to estimate our heart rates, but in different ways. Instead of looking at the voltage over time, \",\n \"We now transform the data to show us what frequencies are present in the continuous signal. \",\n \"In this frequency domain, we now ignore time and consider how much power of each frequency there is in a segment of data. \",\n \"If you place your fingers the same way you did when looking at your ECG, you should start to see a peak. \",\n \"It may help to return to Module 2 and look at the raw data first. \"\n ]}
\n{[\n \"If you missed Module 2, here are the instructions: You can take off the muse from your head and place a finger on your right hand on the muse's reference electrode (in the center of the forehead). \",\n \"We can then place a finger of our left hand on one of the eeg electrodes. Lets use the left forehead electrode (position AF7). \",\n \"So place your left fingers pinching the left forehead electrode, and your right fingers pinching the center electrode. \",\n \"Otherwise try to hold the Muse as still as possible, and relax your body. \" \n ]}
\n{[\n \"In this new plot, Along the horizontal axis is the beats per minute, or the frequency of the peaks in your ECG. \", \n \"The vertical y-axis shows the power of the rhythms in the data at each frequency, or how large the changes are between peak and through of the oscillations. \",\n \"The pink ball shows the estimated peak of the spectra, or your estimated heart rate. \",\n \"In the following experiment, it is only this pink value that will be saved over time while we record. \",\n \"Here is an example of what the plot should look like with a strong heart rate signal. \",\n \"Notice that there are two peaks, 60 BPM and 120 BPM. The larger peak is called the 1st Harmonic, a multiple of the true heart rate. \",\n \"These harmonics are common in the frequency domain: \"\n ]}
\n{[\n \"Clicking this button will immediately record the value of the pink dot above for \",\n Settings.secondsToSave,\n \" seconds. \",\n \"Therefore, make sure the chart above looks clean and you can see a clear peak in the spectra before pressing record. \",\n \"We are going to compare two conditions that show a clear difference in heart rate due to blood pressure changes: \",\n \"Standing and Sitting. \",\n \"You will record two sessions, one standing and one sitting, pick the order randomly but keep track of which output file is which\"\n ]}
\n{[\n \"A .csv file will be saved that can be opened in Google Sheets. \",\n \"Remember for each person to record two files, one while standing and one while sitting. \", \n \"Here is an example of what the data will look like once loaded. \",\n \"The first column shows the time in msec of each estimate of heart rate. The second column shows the heart rate estimate in BPM. \",\n \"Each row represents an estimate from the previous 8 seconds of data. \",\n \"This is because you need a long segment of time to estimate frequency of rhythms over time. \",\n \"Each subsequent row is from an 8 second chunk of data, but is taken about 400 ms after the previous row. \",\n \"You can see the time values increase about 10000 ms during the recording, representing the 10 seconds of data. \",\n \"So 10000 milliseconds divided into ~400 ms shifts per row gives us the rough number of rows (~25). \", \n \"The graph shows the values plotted over time, showing that the 1st Harmonic value of 120 BPM was incorrectly recorded on a few windows. \"\n ]}
\n{[\n \"In this module the analysis will be much easier, since much of the work has been done by the webpage with the fourier transform and peak detection. \",\n \"The following youtube video will show you how to open the file in Google Sheets, rename it, plot the data, remove any harmonics or outliers, \",\n \"then take the average heart rate over the window as an estimate of your heart rate. \"\n ]} \n \n\n Link to example google sheet from video. \n \n
\n{[\n \"Finally each of you will enter this estimated heart rate for both sitting and standing into a new anonymized google sheet that we are sharing as a class. \", \n \"We will use this shared google sheet which combines all our data in order to compute group statistics and to compare the two methods of estimating our heart rate. \" \n ]}
\n\n\n Your data is currently recording, \n once complete it will be downloaded as a .csv file \n and can be opened with your favorite spreadsheet program. \n Close this window once the download completes.\n
\n\n {specificTranslations.intro1}\n
\n\n {specificTranslations.intro2}\n
\n\n {specificTranslations.neurons1}\n
\n\n {specificTranslations.neurons2}\n
\n\n {specificTranslations.neurons3}\n
\n\n {specificTranslations.oscillations1}\n
\n\n {specificTranslations.oscillations2}\n
\n\n {specificTranslations.hardware1}\n
\n
\n {specificTranslations.hardware2}\n
\n {specificTranslations.hardware3}\n
\n
\n {specificTranslations.hardware4}\n
\n \n Image Source - Wikipedia \n
\n
\n {specificTranslations.hardware5}\n
\n {specificTranslations.muse1}\n
\n
\n \n
\n
\n Image Source - @urish \n
\n
\n {specificTranslations.muse2}\n
\n \n
\n \n
\n
\n Image Source - EEG101 \n
\n
\n {specificTranslations.muse3}\n
\n {specificTranslations.signal1}\n
\n \n
\n Image Source - EEG101 \n
\n
\n
\n {specificTranslations.signal2}\n
\n\n {specificTranslations.credits1}\n NeurotechEdu. \n
\n\n {specificTranslations.credits2}\n Interaxon. \n
\n\n {specificTranslations.credits3}\n muse-js \n {specificTranslations.credits4}\n A Techy's Introduction to Neuroscience. \n
\n\n {specificTranslations.credits5}\n eeg-pipes \n {specificTranslations.credits6}\n Muse 2016 Headband + Web Bluetooth.\n
\n{specificTranslations.description}
\n{'Click or tap on the rectangle to toggle sound'}
\n{[\n \"Press connect above to see the chart.\" \n ]} \n
\n{\n \"Last module we recorded the frequency bands while people had their eyes open or closed for 10 seconds. Although we predicted there should be larger alpha power when the eyes are closed, are results showed the opposite effect:\"\n }
\n{specificTranslations.description}
\n{[\n \"If you have not already, it is now time to place the muse on your head. \",\n \"Getting a good connection between the device and your head is crucial. \",\n \"The following video gives tips on how to get a great connection: \"\n ]}
\n{[\n \"The above chart shows the data coming from each of the four eeg electrodes. \",\n \"Time, again, is on the horizontal X-axis and voltage is on the Y-axis. \",\n \"An offset has been added to the electrodes to separate them vertically. \",\n \"The saturation of the lines is controlled by the amount of noise in the data. \",\n \"That is, the cleaner the data, the more rich the colours will become. \",\n \"Conversely, when the line gets dim, there is too much noise present in the signal. \", \n \"I have set the filter settings above very loose (.1 to 100 Hz) to let in as much noise as possible so we can learn, we will restrict them in future modules. \"\n ]}
\n{[\n \"Before we try to use the EEG to estimate brain activity, we first need to observe what other things can influence the signal. \",\n \"Artifacts refer to non-EEG noise in the EEG recording that will cloud the results we want from the brain. \",\n \"There are many sources of noise that the EEG can pickup because all it is doing is recording the voltage changes on the head. \",\n \"Any other source of voltage or change in resistance of the sensor can affect the signal. \",\n \"Here is a great overview of the range of possible EEG artifacts that you might come accross. \" \n ]}
\n{[\n \"Your assignment this week will be to record plots of four different types of EEG artifact. \",\n \"I will show you examples that I made using screenshots of the plot above. \", \n \"To take a screenshot on a mac, press (βCommand + β§Shift + 4) and select the area of the screen you want to save. \",\n \"You can also record data into a .csv file below and plot artifacts using Google Sheets. \", \n \"After making your plots, play around with the filter settings above (cutoff Low and cutoff High) and observe how they change the plotted data. \", \n \"How can we use these filters to help collect cleaner data free from some of these artifacts?\"\n ]}
\n{[\n \"The eye balls create an electrical field, as they move around they create electrical potentials that are picked up by the EEG sensors. \",\n \"Also, the eye lids, as the pass over the eye, create an electrical field that is picked up by the EEG electrodes. \", \n \"Eye movement and Blink artifacts are some of the largest and most difficult to remove from the data. \", \n \"The best way to get clean data is to try to limit blinking and eye movements. \",\n \"These eye movement signals can also be treated as a signal of interest, called the Electrooculogram (EOG)\"\n ]}
\n{\"Eye Blinks: \"}
\n{\"Horizontal Eye Movements: \"}
\n{\"Vertical Eye Movements: \"}
\n{[\n \"When signals are sent from our brain to our muscles the motor neurons release Acetylcholine onto the muscle fibres. \",\n \"This release causes the muscles to contract, and these contractions create electrical potentials that are also picked up by electrodes. \",\n \"This signal is called the Electromyogram (EMG), and can be measured over any muscle on your body. \", \n \"EEG electrodes therefore pickup muscle signals from your face muscles when you smile or frown, and from your chewing muscles. \", \n \"You should make sure to not be chewing gum for this reason while recording EEG. \"\n ]}
\n{\"Example of Jaw Clenching: \"}
\n{\"Zoomed in chewing (notice the difference horizontal axis time range): \"}
\n{[\n \"The EEG electrodes measure the voltage of the human body between two points. Voltage, according to Ohms' law, \",\n \"is modified both by the current and by the resistance (V = IR). \", \n \"Therefore, if current stays the same, changes in the electrical resistance of the electrode-body connection can change the voltage. \", \n \"These changes in voltage will not be distinguishable from real changes in voltage from the brain. \",\n \"This is one reason that mobile recording of EEG is difficult, movement of the electrodes against the head leads to large voltage changes not due to brain activity. \", \n \"Therefore we want the muse to be firmly secured to the head, as tight as possible, and we want to avoid excessive movement during recording. \"\n ]}
\n{\"Tapping on the muse moves the electrodes against the head: \"}
\n{[\n \"Resistance between the sensor and the body leads to less current and smaller measured voltage. \", \n \"While out salty wet skin conducts electricity well, the dry skin and air does not, nor do oils and dirt that our skin is covered in. \",\n \"Because voltage changes as resistance changes, we can sometimes see slow drifts in our EEG data as the connection strength changes. \",\n \"One example is the moment you place the muse on your head, as the resistance decreases the voltage decreases as well as you can see here. \", \n \"These slow drifts can also occur due to sweating, which decreases the resistance slowly over time between the head and the sensor. \"\n ]}
\n{\"When you first put on the muse or when you sweat long drifts occur: \"}
\n{[\n \"Finally, there is electrical noise all around you. Buildings are supplied with 120 Volts of alternating current to power our lights and electronics. \",\n \"This electrical alternates in polarity 60 times a second, allowing for much more efficient transportation accross distance locations. \", \n \"Most of our modern electronics convert this AC electrity to DC power (usually around 5 Volts and 2 Amps). \", \n \"All the computers in the room take in AC power from the building and convert it to DC in their power supply. \",\n \"Therefore, there is alot of 60 Hz electrical noise in any building. \", \n \"Most EEG systems have filters to remove some of this noise, but even still, it is so strong that it is easily picked up by EEG electrodes. \", \n \"We can use filtering to remove this type of noise from our data. \"\n ]}
\n{\"Electical noise at 60Hz from the AC power in the room (I held my wired headphones up to one side then the other of my head): \"}
\n{\"Zoomed in 60-Hz electrical noise (notice the difference horizontal axis time range): \"}
\n\n {\"When you are recording raw data it is recommended you \"}\n {\"first set the sampling point between epochs to 1, then set the epoch duration to 1. \"}\n {\"Once the live chart disappears entirely you have done it correctly. \"}\n {\"This will make it so every row of the output file is a single time point and make the data much easier to work with.\"}\n
\n\n Your data is currently recording, \n once complete it will be downloaded as a .csv file \n and can be opened with your favorite spreadsheet program. \n Close this window once the download completes.\n
\n{'Connect the device above to see the plot'}
\n{specificTranslations.description}
\n{[\n \"In module 3 we saw how we can use a mathematical technique called the fourier transform to estimate what frequency is present in the ECG data, to estimate heart rate. \",\n \"A fourier transform turns any series of numbers into a summed set of sine waves of different sizes. \",\n \"For review, the following animation shows how a single time-series of data can be thought of as the sum of different frequencies of sine waves, each of a different magnitude. \", \n \"The blue bar chart at the end of the animation shows what is called the frequency spectra, and indicates the power at each frequency.\" \n ]}
\n{[\n \"The Fourier transform is a mathematical technique originally developed in the early 1800s in order to mathematically model the movement of heat. \",\n \"A discrete fourier transform (DFT) is this mathematical technique applied to digital data (a function sampled at different time points), like EEG data. \",\n \"In order to compute this efficiently, a pair of psychologists and statisticians created the FAST fourier transform (FFT) in the 60's, during the cold war \",\n \"as a signal processing technique needed to triangulate possible Soviet nuclear launches from a hypothetical array of sensors around the Soviet Union. \",\n \"The FFT has gone on to be one of the most useful and used algorithms ever created, and is used in a wide array of digital tools. \", \n ]} \n \n Follow this link to an excellent interactive tutorial on the graphical understanding of the fourier transforms using sound, drawings, and images\n \n
\n{[\n \"We will not cover the mathematical formula or algorithm behind the DFT or FFT here. \",\n \"A fourier transform turns any series of numbers into a summed set of sine waves of different sizes. \",\n \"The following animation shows how a single time-series of data, can be thought of as the sum of different frequencies of sine waves, each of a different magnitude. \", \n \"Amazingly, before digital computers in the late 1800's, in order to model details of light movement, a physical machine was constructed to perform these calculations by hand. \",\n \"This amazing Harmonic Analyzer still provides an excellent intuition into the mechanics behind decomposing a time series signal into a sum of sin waves. \" \n ]}
\n{[\n \"Most computer programming languages now provide an easy way to compute the FFT on time series data. \",\n \"Other methods of converting time series data into the frequency domain also exist, such as wavelet analysis, in which the product of the data and a family of different frequency wavelets is used to esitmate the data decomposition. \",\n \"The resolution in frequency of the FFT depends on the NUMBER OF TIME POINTS. \",\n \"The range of frequencies provided by the FFT depends on the sampling rate of the data , in our case 256 Hz provides frequencies up to 128 Hz (half). \" \n ]}
\n{[\n \"Importantly, the uncertainty principle applies to the FFT as well. \",\n \"It is impossible to know both exactly when something happens, and what frequency it happens at, at the same time. \",\n \"As an intuition, imagine that in order to estimate your heart rate you obviously need more than a single time point of data, you need multiple beats of the heart (at least one). \",\n \"Therefore, the resultant estimate of heart rate will not be precise in time, and will apply to the entire time window you put into the FFT. \" \n ]}
\n\n{[\n \"In module 4 we plotted the range of various artifacts that are common in EEG data. \",\n \"Obviously, these artifacts remain (if not filtered out) in the window of data that is used to compute the FFT. \",\n \"Therefore, NOT ALL THE DATA IN THE SPECTRA IS FROM THE BRAIN!!! \", \n \"For instance, blinks show up as low frequencies, and you have already seen how heart artifacts in the data would show up as low frequency power as well, but these are not brain activity. \",\n \"For your assignment this module, for each type of artifact you observed in Module 4, now you will observe what the spectra looks like, take a screenshot of the result and save these in a google doc. \",\n \"To take a screenshot on a mac, press (βCommand + β§Shift + 4) and select the area of the screen you want to save.\"\n ]}
\n{[\n \"Press the following button after adjusting the settings above in order to record the live spectra over time into a .csv file. \"\n ]}
\n{[\n \"A .csv file will be saved that can be opened in Google Sheets. \",\n \"Here is an example of what the data will look like once loaded. \",\n \"The first column shows the time in msec of each estimate of the spectra.\",\n \"Each row therefore represents an estimate from the previous 8 seconds of data EEG data, and the windows used to compute each row are overlapping. \",\n \"Again, this is because you need a long segment of time to estimate frequency of rhythms over time. \",\n \"You can see the time values increase about 10000 ms during the recording, representing the 10 seconds of data. \",\n \"So 10000 milliseconds divided into ~400 ms shifts per row gives us the rough number of rows (~25). \"\n ]}
\n{[\n \"The spectra are then shown on each row. Each column represents the power at a different frequency. \",\n \"The first row shows the frequency and channel label for all the data in that column. \",\n \"So the first 30 columns after the timestamp are the 30 frequencies from the TP9 electrode \",\n \"(where 30 is the number of frequencies saved which can be adjusted with the FFT Slice Settings). \",\n \"After that, the next electrode starts, with another 30 frequencies.\",\n \"After columns for all 30 frequencies from all four electrodes, another 30 columns show zeros, this is for an optional auxillary channel we are not using here. \",\n \"Finally columns are saved to record the exact frequencies of each bin of the FFT (redundant with the column names). \"\n ]}
\n{[\n \"The second part of the assignment for this module involves parsing this output file to organize the data into a format that can be plotted like the live plot on the page. \",\n \"Data will be averaged over time, and then data for each electrode will be organized and used to make a chart of the spectra\"\n ]} \n \n\n Link to example google sheet from video. \n \n
\n\n Your data is currently recording, \n once complete it will be downloaded as a .csv file \n and can be opened with your favorite spreadsheet program. \n Close this window once the download completes.\n
\n{specificTranslations.description}
\n{specificTranslations.description}
\n\n Your data is currently recording, \n once complete it will be downloaded as a .csv file \n and can be opened with your favorite spreadsheet program. \n Close this window once the download completes.\n
\n\n Your data is currently recording, \n once complete it will be downloaded as a .csv file \n and can be opened with your favorite spreadsheet program. \n Close this window once the download completes.\n
\n