Consistent Best Effort


Posted on 4th April 2008 by Gordon Johnson in Uncategorized

, , , , , , ,

This week I have been discussing the basic principles of neuropsychological assessment, and its two foundational assumptions: the ability to reconstruct pre-morbid IQ and the need for “consistent best effort”. Yesterday’s blog dealt with the pre-morbid IQ. Today, we will discuss the issue of “consistent best effort.”

The number side of the neuropsychological assessment is based upon the theory that a neuropsychologist can make certain conclussions about pathology based upon an examination of the pattern of test scores. The process of doing this is called “discrepancy analysis”, meaning that if there is a discrepancy in certain areas, this points to pathology. Two other terms are important: “relative weakness” and “intraindividual comparison”. If while doing the intraindividual comparision (mean comparing the patient, only to his or her own scores versus the population as a whole) a “relative weakness” shows up, then that means something.

In a perfect world, it is a beautiful theory. You chart the scores, the “relative weakness” jumps out at the neuropsychologist, you look to the part of the brain that controls that area of function, and thus, make a diagnosis. The fundamental problem is that you must be able to presume that the test subject was making the same effort during the test where he or she did poorly, as across the entire battery of tests. But can we make that assumption?

I like to quote from depositions I have done to make these type of points, and I will do that again. My apologies to my son for my references to his middle school running career.

12 Q (By Mr. Johnson) Do you still have your Exhibit Number 1
13 before you?
14 A I do.
15 Q Page 6?
16 A Yes.
17 Q Now, as I understand what you’re saying in the first
18 paragraph of Page 6, what you’re saying is that because you
19 cannot be sure that the patient did not give optimum effort,
20 that you can’t reach conclusions based on the data in those
21 testing — in that testing; is that correct?
22 A I can make certain conclusions, but not on her current
23 status, on that date. That’s what I’m — all I’m trying to say
24 is this set of data had serious reservations because of lack of
25 effort.

1 Q Now, there are any number of things — strike that. Let’s
2 talk about the continuum of effort when you’re giving someone a
3 test; all right? I’ll give you an example.
4 My son, who is a 13 year old, goes out and runs a six-
5 minute mile, and he gave better effort than anyone else in the
6 class if you judge it just based on his performance, because he
7 won the race; okay?
8 A Got you.
9 Q Now, would that be considered best effort?
10 A It was certainly a sufficient effort to be recorded, yes.
11 Q Two months later in a track meet in his conference meet,
12 he’s able to run a five-minute, six-second mile without
13 significant change in this training status. In comparison to
14 the gym class — in comparison to the conference meet time of
15 five minutes and six seconds, did he give best effort in gym
16 class?
17 A There are other variables that have to be considered, and
18 I’d have to know other things. I’m not really following you.
19 Q Okay. Tell me what the variables would be.
20 A Like the environmental conditions, the contingencies if he
21 won or if he didn’t win, the particular mood or attitude that he
22 had on that day, how his physical health was, if he had a cold,
23 if he had some sort of limitation.
24 Q Now, we always have all of those limitations anytime we
25 give someone any type of test; is that correct?

1 A Exactly right.
2 Q If we were going to pick an example of when we might get
3 the highest percentage of people giving maximal effort or
4 optimal effort, is there a better example than the law school
5 admission test?
6 A Well, I’ve never seen the law school admission test, but if
7 it’s like the test that I took to get to graduate school, then
8 one certainly has to do well, as best as they can, yes.
9 Q And can we — if there ever — can we ever presume a higher
10 likelihood of maximum effort in an academic test than we would
11 in something like a law school or a medical college admissions?
12 A Well, I agree. I mean, one can’t do better than one can
13 do.
14 Q But what’s unique about the law school and the medical
15 school admission test, is people’s whole lives revolve around
16 how they do on this test; correct?
17 A Well, that’s probably their interpretation, but it’s not
18 real. They probably think —
19 Q And that thinking that would convince them at least
20 relative to other variables to give it their best shot?
21 A I would think so, yes.
22 Q Despite that, sometimes people who are testing in high-
23 pressure situations like a law school admissions test or a
24 medical college entrance exam, do not wind up at their optimum
25 performance level; correct?

1 A I presume that’s correct.
2 Q And what explanations for that would do?
3 A Again, we just went through some of them. They have a
4 cold, they’re worried about money, they have stress at home,
5 they have stress on the job, I mean, there are all kinds of
6 events that could influence particular effort on a particular
7 day.
8 Q Or actually the stress of the test itself?
9 A Well, yes, of course. There’s some people who don’t do
10 well on tests.
11 Q And there are some people who do worse the more the
12 pressure is?
13 A Right. It’s not really the pressure; it’s how the patient
14 manages the pressure that’s the issue.

Now as we consider this long introduction in the context of the search for “relative weaknesses”, what does that mean? What if our test subject was only using the gym class effort level, versus the conference meet effort level? Can we make statistical comparisons then? Or should we compare that performance to how people do in gym class, and not comparing how they do in more stimulating environments?

Neuropsychology is a science, right? They should have control out all of these variables, right? Guess again, not because they don’t want to, but because they are dealing with human beings, and in brain injury evaluations, human beings who prevented from doing what they are presumed to do, based upon the precise disability for which we are evaluating them: brain damage.

Next: The Scope of the Problem for Brain Injured Person in Giving Consistent Best Effort.

Best Performance Method in Neuropsychological Assessment


Posted on 3rd April 2008 by Gordon Johnson in Uncategorized

, , , , , , ,

There are two fundamental premises upon which the statistical application of the science of neuropsychology is based: The first is that a determination can be made of what a given individual’s premorbid abilities were. The second is that an individual is giving consistent best effort throughout the test battery. Neither assumption works perfectly, but the extent to which these two assumptions work well enough, will determine whether legitimate statistical based diagnostic conclusions can be incorporated into the assessment. In today’s blog, we will discuss the Premorbid Ability assumption. In tomorrow’s, the consistent best effort issue.

Premorbid ability. By premorbid ability, we mean a given individuals abilities prior to the onset of the accident or disease process. If no assumptions can be made about premorbid ability, no diagnosis about “cause” can be made by a neuropsychologist. All they are capable of saying is that a given individual has certain weaknesses and disabilities, but no definitive diagnosis can be made. Thus, some method of assessing premorbid ability is essential.

Most neuropsychologists don’t look at enough information in determining pre-morbid IQ. They base far too much of their assessment with respect to pre-morbid ability on the test battery itself. In our earlier example of the person with the IQ of 135 post the accident, that is less of a problem. Clearly a person who has a post-morbid IQ of 135, was very superior before the onset. But most cases are not so clear cut. A previously brilliant person may not continue to have a very superior IQ after the accident. If certain deficits bring the person down into an IQ range of 110 or so, we would likely need to look for other evidence to determine IQ.

One way is by looking at the areas where they still have strengths. If their average scores are in the 130 or above area, and there are a few scores in areas we might suspect would be effected by the injury, then it might be easy to say this person was very superior before. But again, that is the easy pattern to spot. Most profiles are not that obvious.

Another method is to look at certain subtest scores, where it is believed that a given ability is unlikely to be substantially effected by the given injury. Reading scores are often thought to be an ability that is rarely changed significantly by a mild or moderate injury. Thus, a neuropsychologist might say that a person with a “very superior” reading score and a much lower current IQ, had pathological deficits, based on the retained ability to read at a high level.

All of these methods work far better with someone with a very high IQ. When you are dealing with people in the average range, IQ’s of 90-110, it becomes much more difficult to make such assumptions about premorbid IQ from subtest scores.

Another method is to assume IQ based on a assessment of that person’s educational level. So a person with a college degree would be assumed to have a higher IQ than someone without. The obvious flaw in such logic, that some brilliant people don’t go to college, isn’t even the most significant problem. The significant problem is that it groups all college graduates together. Ask anyone who went to college. Not all of their classmates were of equal intelligence and ability.

Another method, one I believe to be considerably better than the first two, is called the ‘best performance method.” The best performance method is based upon the assumption that a person’s highest areas of achievement are the best indicators of premorbid ability. If these areas of highest achievement are in contrast to significantly lower subtests scores that may point to pathology.

Of course, there is considerable disagreement as to how to apply the “best performance method.” Many neuropsychologists dismiss it as they interpret this method as applying only to the best performance on individual tests, within the full battery of tests. That would mean if the person got 99% in arithmetic or vocabulary, that would mean that such person is in the 99%. It is easy to poke holes in a restricted use of the “best performance method” because we all have normal variances in what we are good at.

However, another interpretation of the best performance method is that it makes a full assessment – not just of the scores on the given battery of tests – but also the person’s real world performances or achievements. For example, if a person has graduated from medical school, one assumes that they are very near the top of the pre-morbid ability level. Likewise, if they have risen to the top of any profession, they would be assumed to be near the top.

In my opinion, the overall preferred method, which of course is harder to reduce to statistical probabilities, is to use of the real world “best performance method”. Such method considersall factors, school records, work performance records, areas of retained strength on the test. If someone got a math score of 700 and a verbal score of 700 on the SAT when applying to college, they clearly were way above average at that time. If they went on to graduate from a competitive law school or medical school, we must almost assume that they were at the superior or likely very superior level.

If the scores were good, but not great, if they graduated from college with more than a B average and went on to have a successful career, we can’t assume they were only average. Whether they are high average or superior is open to interpretation but that is what professionals are supposed to do: make subjective interpretations of complex multi-faceted variables, to reach conclusions.

Who a person was before injury is far more complex than how well they do now on a reading score. Only if neuropsychologists look at not the basic outline of a person’s premorbid life, but level of achievement within that life, will neuropsychology be able to identify the true areas of acquired deficits and disability.

Tomorrow the concept of “consistent best effort.”

Dilated Perivascular Spaces in Identifying Mild Brain Injury

1 comment

Posted on 26th March 2008 by Gordon Johnson in Uncategorized

, , , , ,

We began our series on “Advances in Neuroimaging” with yesterday’s blog on Increased Field Strength – the improvement from the 1.5 Tesla MRI’s to the 3 Tesla MRI’s. Today’s blog will discuss the evolving research that seemingly insignificant evidence of abnormalities on lower field strength scans, can illuminate evidence of traumatic injury. This is particularly true of what is technically called dilated perivascular spaces, also called Virchow Robbin Spaces. In essence, these are areas surrounding blood vessels in the brain, where myelin sheath or other brain tissue is missing, and thus show up as holes (white spots – also called UBO’s – unidentified bright objects) on the MRI scans. The technical term in an MRI report would be “areas of abnormal increased signal intensity.” Myelin sheath is the insulation type substance that protects the length of most multilevel axons in the brain.

Even though the 3T scans allow us to see these bright spots much clearer, this one term, “dilated perivascular spaces”, is still being used to describe lots of very different types of pathology. Again, excerpts from a recent deposition I conducted:

10 Q. We have the term called dilated
11 perivascular spaces that we’ve talked about.
12 Peri meaning?
13 A. Around the vessels. It’s the
14 spaces around blood vessels of the brain.
15 Q. And in a normal brain, what is in
16 those — why are there no spaces?
17 A. Well, there actually are spaces.
18 There’s spaces in everybody. It’s just that
19 they’re very, very small. And in some
20 patients you really see very few, or you
21 don’t really see hardly any, but they’re
22 there.
23 And then if they enlarge, because
24 you have lost substance in the brain around
25 them, then you refer to them as dilated
1 perivascular spaces, or enlarged spaces. And
2 what they fill in with is basically water;
3 cerebrospinal fluid is water.
4 Q. What type of brain matter is lost
5 in these, in the dilated perivascular
6 situation?
7 A. It’s basically a white matter
8 substance of the brain predominantly, because
9 that’s where you — that’s where these
10 perivascular spaces tend to be is mostly in
11 the white matter. So basically what you’ve
12 lost is some of the connecting fibers.
13 Q. Now, white matter is the axonal
14 part of the brain; is that fair?
15 A. Right. It’s the connecting fibers
16 of the brain. I made the analogy earlier of
17 the telephone systems. Telephone systems on
18 the surface of the brain, they’re basically
19 neurons. The connecting fibers are the
20 axons. And those connecting fibers are what
21 make up the white matter.
22 Q. And they call it white matter
23 because it’s white when you autopsy the
24 brain?
25 A. Depending on how you fix the
1 brain, yes.
2 Q. What is it that you’re seeing
3 that’s white? Is it the axons themselves or
4 the insulation around it?
5 A. It’s the insulation around them,
6 the myelins.
7 Q. And when we see dilated
8 perivascular spaces, are we seeing absence of
9 axons or absence of the insulation?
10 A. It could be both. We’re seeing an
11 absence of one or the other.
12 Q. Is the insulation considerably
13 larger in scale than the axons are?
14 A. Yes.
15 Q. Do you have any sense of the
16 magnitude of the difference?
17 A. An axon is on the order of about
18 50 microns. And then it has a mild sheath
19 around it. So that covering around that. So
20 the whole thing is really small.
21 Q. How small is a micron relative to
22 a millimeter?
23 A. It’s a thousandth of a millimeter.
24 Really tiny.
25 Q. So 50 microns would be 1/20th of a
1 millimeter?
2 A. Pretty small.
3 Q. Can you see axons in a human
4 macroscopically?
5 A. Only collections of them. Only
6 bundles of them. Large groups of them. You
7 can’t see 50 micron scales.
8 Q. And relative to the 3-Tessla MRI,
9 what is its resolution in terms of pathology
10 and the smallest pathology you see?
11 A. I think that depends on the type
12 of pathology. I would say in general you’re
13 in the 1 millimeter resolution range.
14 Depending on the pathology, you could go
15 smaller. Some pathology you might have to go
16 a little bigger. I feel very confident
17 calling 1 millimeter lesions.
18 Q. How many axons grouped together do
19 you think you would have to see to be able
20 to see it on the MRI?
21 A. Well, at the very least hundreds,
22 and probably thousands.
23 Q. Some are thicker than others?
24 A. Right.

For a more detailed explanation of the pathology of diffuse axonal injury, see

A dilated perivascular space on a 3T scan is no longer a vague bright dot but now has definition, measurable size and distinguishable shape. A neuroradiologist may be able to distinguish between such dilated spaces that can be caused by trauma, from other disease processes. .

25 Q. Is there weighting that goes into
1 your differential diagnosis when you look at
2 a dilated perivascular space in terms of more
3 likely trauma, more likely aging, more likely
4 microvascular? Is there — can you look at
5 the character in relation to the location of
6 perivascular spaces and shift a probability of
7 one diagnosis versus another?
10 THE WITNESS: Well, one of the
11 things that I’m looking at right now are
12 different types of perivascular spaces. And
13 we have a study that we’re conducting where
14 we’re looking at — I actually think there
15 are two types of perivascular spaces. There
16 are perivascular spaces that I would refer to
17 pathologic, and perivascular spaces that would
18 I would consider developmental.
19 So I’m going to eliminate the ones
20 that have — and what you’re asking me I’m
21 going to separate out the developmental ones.
22 The developmental ones, I think are
23 very round. They’re usually in the deep
24 basal ganglion region of the brain. They’re
25 very common. They can be extremely large.
1 They’ve been described in literature to be
2 well over a centimeter in size. Very big.
3 But I don’t think they ha ve any clinical
4 significance at all.
5 Then there are dilated perivascular
6 spaces that I think are pathologic, meaning
7 that something caused them, whether that is
8 aging, whether that is a disease process,
9 whether that’s trauma.
10 I think that differentiating
11 between those requires you to look at a
12 variety of factors. And that is, does the
13 patient have any other disease condition?
14 What is the age of the patient? What are
15 the size of these perivascular spaces relative
16 to the age of the patient? Are they greater
17 than you anticipate for that patient’s age?
18 Is the location predominately in areas of the
19 brain where those particular disease processes
20 are most common? Do they fit together in that
21 way?
22 So my opinion is that we probably
23 will be able to over time improve our
24 differential diagnosis. I think we can put
25 it into two categories right now. And then
1 I think beyond that it really requires
2 correlating it with clinical information, the
3 age, and the locati of the perivascular
4 spaces.

In summary, we have now covered improved field strength and dilated perivascular spaces. In tomorrow’s blog, we will address the need for tailored protocols in properly investigating Mild Brain Injury and the existence of Post Concussion Syndrome, aka, Subtle Brain Injury.