BME319 Ocular Motility Lab Report Frequently Asked Questions.
Last modification: 17 February 2010, 11:55
--------------------------------------------------------------------------------
These are questions we have received from students in previous years. There are bound to be some repetitions listed here. Please read through them before to see if your question has already been asked and answered.
(Names have been removed to protect the guilty. They know who they are...)
--------------------------------------------------------------------------------
Q: This is kind of a trivial question, but I was wondering if you wanted
standard deviation or standard error for error bars in the plots.
A: We are not asking for them, but feel free to add them.
---
Q: In #6, are the two conditions moving and stationary (for imaginary
only), or are the 2 conditions imaginary and non-imagining with both
moving and stationary (i.e., 4 conditions to compare)? Could you please
clarify exactly which conditions we need to compare.
A: Compare and contrast the differences between moving and stationary for the real and imagined cases.
Q: Do we have to use the trials where the subject was counting in
the dark for any part of the write-up?
A: The "counting" trials are to illustrate the performance of the VOR when the subject is distracted or not paying attention. What do you see in this data? How do these results compare to the other cases?
---
Q: In order to analyze the data from the 2 rotation in the dark trials (subject
was imagining stationary and moving targets)...am I measuring the slope of
the smooth parts of the velocity tracing for one eye? I am confused on how
I use the chair calibration data. Thank you for your time!
A: You measure directly from the chair velocity channel. That's why we provided you with the chair calibration data; it shows you what 20 degrees/second looks like for comparison with the chair velocity recorded during the trials.
You can determine the eye velocity by reading it directly off of the paper, although this can be a problem if the channel is noisy. Alternatively, you can calculate the eye velocity as the slope of the position data. Since the chair was moving (approximately) sinusoidally, the velocity is always changing, so be sure to compare the eye and chair velocities at the same time (which you can do because VOR is pretty close to instantaneous, with a delay of less than 20 milliseconds).
---
Q: On question 6, you are asking us to compute averages
of the peak VOR eye velocity for two rotation in the dark conditions,
I was not sure if we need to use the position graph of the eye, or the velocity
graph from the sheets we have forÊcomputing eye velocity.
A: You can always get the peak velocity from the position trace if, for example, the velocity trace was too noisy. I do not recall that any group had problems with the chair or eye velocity noise this week, so you should probably be able to simply use the velocity channels. If one eye is noisy, use the other one.
Q: That works for the position tracing, but how does that help with the velocity tracing unless you already know the velocity at a given point?
A: distance=velocity/time. As for time, you know the paper speed and you have the 1-second-interval tick marks in the trace ABOVE the target trace.
---
Q: I don't understand how to use the feedback equation to predict what feedback should be needed to sustain oscillations. Are there any hints that you can give me about that? And does saccadic oscillations mean that successive saccades should go in different directions, or is just enough to have only saccadic motion?
A: Add the external feedback as supplied by the computer experiments to the internal feedback built into our brains (-1) to get the TOTAL feedback. But I think the easier way you can predict what level is necessary is to work out an ideal experiment as follows: when the eye moves x degrees in one direction, the target will move y degrees in either that direction or the other, depending on the sign of the external feedback. y depends on the magnitude of the feedback. It's simpler to do than to explain, so let's work an example. Ready? Here goes: Since the eye starts at zero, and then must meet the target's initial position at 5 degrees, work out manually what will happen for different gains. If the gain is +0.25, when the eye jumps from 0 to 5 deg, the target will be driven in the SAME direction by 1/4 that distance and will go from 5 to 5+(0.25*5)=6.25 degrees. Since the target moved, the eye is STILL off-target so it must try again. Repeat this process a few times to see what happens. Try it for each gain (pos and neg) and see what you come up with.
Saccadic oscillation, like any other oscillation, means that it is centered around some point, therefore they must be oppositely directed. Otherwise, what you have is a run of saccadic intrusions. (you can see these in some of the data taken over the three days -- i pointed it out whenever it happened in class; if you were lucky, you might have been near the computer at just the right time.)
---
Q: In the instruction packet you sent out to us, I am seeing a strange
character (it looks like a combination of a T with a superscript Y) on page
9. I was wondering if you could clarify what this letter should have been.
It's in Point 1 right after the underlined text in the parentheses (i.e., G
vs. ??). I'd like to know what I should be plotting G versus. Judging by
the context, I should be plotting G vs +-V. Is this correct?
A: It's "T-dot", i.e target velocity.
---
Q: I am getting the right duration for each cycle so I am confident my time
base is fine. I did check the amplitude and the peak to peak amplitude was
15 squares. 15 squares equals 15 degrees with the calibration technique we
used, correct? Can I use my calculated values for the target velocity
instead of the ones the computer was supposed to have generated or is there
something I could be overlooking still? Thanks for any suggestions you can
offer.
A: Two things to check before panicking: First, are you remembering that
the target moved from -15 to +15, giving a p-p amplitude of 30
degrees? Second, if you are uncertain of the timebase settings, look
at the very top trace (above the target trace). The marks made there
are one second apart.
The p-p is 30 degrees. This suggests that one square has to be 2 degrees.
---
Q: i have a problem finding the phase lead and phase lag for the sinosoidal
target (question 2). Can you please tell me how to measure these??
A: A complete cycle (e.g. from -15 to +15 back to -15) is 360 degrees. This takes a given amount of time, which you can measure. Find a point on the target tracing, and its analogous point (e.g. rightward peak) on the eye tracing. How much time is between them? The answer should be clear from this point on.
---
Q: 2. This expression is confusing: "Plot the phase lag in degrees, of eye
velocity..." Isn't this a contradiction? How do you find velocity in
degrees - it should be degrees per second.
A: It's not a contradiction, but I agree that you can read too much into it. Simply plot the phase lag against the frequency of the stimulus.
Q: I'm thinking that to find the lag or lead, I should just measure the target
and eye positions and then subtract them. If eye pos. is greater than
target pos. it's a lead. It's a lag if eye pos. is less than target pos.
Is that right?
A: Yes.
Q: Also, to find the peak velocity: Doesn't the ratio of peak velocities (for
the eye and the target) equal the ratio of peak positions since the
derivative of sin is cos and they both have the same amplitude?
A: You need to calculate the velocities.
Q: 4. This questions is really confusing. How do I pin-point the exact
number for the feedback?Ê
In our data, what is feedback internal?Ê
In our data, what is feedback external?
Also, what do I want feedback total to equal and why?
A: Internal feedback, as explained in the handout and post-lab is the visual path. When the eye moves right, for example, an fixed target would move in the opposite direction at the same velocity. Now when we play tricks with the equipment, we can add (+ or -) gain to this system. The total feedback is the combination of the two.
Q: 5. What do you mean by the "...single jump programmed by the saccadic system"?
Also, what is the "continous or sampled-data model"?
A: When the two stimuli are separated by less than a certain duration, one will be ignored and the eye will only make one saccade. Find this duration.
---
Q: can you tell me what is meant by continuous and sampled data model for the saccadic system.
A: As the names suggest, this describes the times during when new inputs can be seen and used to influence new behavior.
---
Q: i have problems finding all the references mentioned in the hand out. i did find four of them. is it okay if i look at similar experiments to find out more??
A: That should be okay, just mention what you referenced.
Q: also in the 3rd question how do we find the velocity of the smooth movements???
A: velocity = distance/time. (BUT DO THIS ONLY FOR ANY SMOOTH MOVEMENTS, NOT FOR THE SACCADES) You can measure over short distances which will give you roughly linear segments.
Q: and what do you mean by "value of T which a single jump was more often
programmed by the saccadic system"?
A: When the two stimuli are separated by less than a certain duration, one will be ignored and the eye will only make one saccade. Find this duration.
---
Q: For question 1, if there are multiple saccades, what should I do? I have this question in several parts of the lab report. Should I just average the velocities of each pursuit in each rising edge/falling edge and use that averaged value?
A: To get the most accurate estimate of pursuit, you should look at as many segments of pursuit for each target velocity as you can. Remember to analyze just one eye, and to separate results for leftward and rightward pursuit.
Q: For question 2, are all of the velocities I'm looking for peak velocities? The reason I ask is that in the first paragraph, peak is only mentioned in the Note, and in the second paragraph, it's not mentioned on the target velocity. Also, when finding peak eye velocity in the second part, should I find this by using MatLab computations to compute the peaks from the data vectors? Another question I have is related to an offset in the eye data compared to the target data. How would I measure phase lag and lead with these offsets? I'm not sure what to do, so I was thinking of just not accounting for the offsets in the calculations, and then adding a note in the discussion about seeing offsets and that the data might be affected by this.
A: You should only perform the analyses for peak velocity, i.e., around the times when the sinusoid passes through zero degrees. You can compute the velocity however you feel best, either by frank differentiation of the target and eye velocities, or a good linear estimate. In the presence of an offset of position, you should determine the true zero of the sinusoid by finding the halfway point between rightward and leftward peaks. From that point on, you can calculate phase just as shown in the postlab notes.
Q: For question 6, since I don't have computer data for this section, should I just eyeball the peak eye velocities?
A: In the days before digital data, we had to take our calipers (or just count squares on the stripchard graph paper) and make physical measurements. While a bit more time consuming, it is just about as accurate. If nothing else, doing one set of analyses just like the pioneers did will give you an appreciation for modern convenience...
---
Q: I had a question on how to load the 19 files onto matlab. I am typing
in matlab, load('ASCII 0222707'); but it is saying that it is not there. The
files are documents, is that what they should be?
A: The load command must include the whole path name for the file. This is simpler if you are already in the file's directory. The session notes (included with the data or available from the web page) cover this in great detail, including text you can copy and paste into MATLAB to help get you started.
Q: To get the graphs to calculate velocity, should we load the first four
documents and do as told on the website? After we get these graphs, do
we then use it to calculate the velocity?
A: You calculate velocity in the most basic way: rise divided by run, i.e, y degrees divided by x seconds. You can do this in MATLAB or by making measurements directly off of the stripchart records.