Thread by @michael_hoerger

As an expert in psychological assessment who has testified in court on cognitive assessments I've conducted, people should interpret carefully the new eClinicalMedicine paper on Covid-related cognitive changes.

Quick 25-pt thread⚡️
🧵1/25

The field of clinical psychology has developed, implemented, and evaluated normed broadband tests of cognition for the past 119 years.

These are highly specialized instruments with carefully selected tests to cover the breadth of key areas of cognition.

2/25

The study did not use one of the well-established, normed, broadband tests of cognition. Instead, it used a novel app-based hodgepodge of tests with little empirical history.

People should be very cautious in interpreting results. There is no vast literature on the tests.
3/25

Many seem to frame the finding as a general decline in cognitive skills, or even as an "IQ loss" of 5 or more points.

This is wrong on many levels.
4/25

If you look at their hodgepodge of relatively unvalidated tests, you see really only one test stands out.

It's something called, "Object Memory -- Immediate." Again, this is not a test anyone would use in a clinical or forensic setting. I had to look it up.
5/25

This image of "Object Memory -- Immediate" is from another of the authors' papers.

It's a short-term memory test of how well one can remember an object AND how it was oriented. This is an atypically specialized test to include in a general cognitive battery.
6/25

We see from another figure that the main change was not in recognizing the correct image, but in remembering the orientation.

E.g., Covid didn't impede recognizing the cherry, but did throw off recognizing whether it was the one with leaves on the left side or right.
7/25

So, it seems Covid impairs the specialized skill of "spatial encoding" or putting an orientation into memory.

8/25

BUT it could be something simpler like impaired attention. Unlike in real cognitive assessments used clinically and forensically, these tests do not have a vast literature.

9/25

For clinical cognitive assessments, there are vast literatures for tests, so we can look up the correlates and see, for example, that a particular test is 3 parts encoding, 2 parts attention, and 1 part processing speed (or whatever).

10/25

Without that literature of correlates, the test could be assumed to measure spatial encoding, but it could be attention, processing speed, spatial reasoning, or other processes.

They included other tests that get at some of those, but perhaps THOSE tests were just junk.
11/25

So, let us assume it's a quite dramatic spatial encoding deficit, but also keep in mind, the test could be picking up on different cognitive processes entirely.
12/25

Nobody should frame this as a general loss of cognitive skills, or an IQ loss.

First, it was not general.
13/25

In cognitive assessment, we frame findings as general if performance across all tests is similar (there are many statistical tables for inferring this).

If there are differences across tests, we highlight the differences in skills as quite meaningful.
14/25

So, it's inaccurate to frame this as a loss in general cognitive ability or IQ.

Second, it's also potentially offensive.
15/25

People should be very careful about any statement using the term "cognitive ability" or IQ, as they have been co-opted by a minority of racial eugenicists ("The Bell Curve") who frame differences in terms of genes, rather than differences in education/opportunity/etc.
16/25

People should also be very careful about saying "IQ" or other value-laden language as it's potentially ableist. A lot of the remarks I see are very insensitive.

In sum, "IQ" language is not accurate here and potentially offensive.
17/25

So, let's assume Covid causes a huge decline in spatial encoding (putting directional orientation into short-term memory), what are the implications?
18/25

This could be one of the factors underlying increases in accidents, especially car accidents.
-Remembering someone's blinker as left instead of right and crashing into them
-Turning left instead of right
-Remembering directions wrong, and changing course last minute.
etc.
19/25

Occupations like air traffic control would seem particularly problematic.

But there are other areas.
20/25

Examples:
-Language learning (b vs d)
-Arithmetic/accounting/finance (+ vs x, 6 vs 9)
-Operating novel/complex equipment & machines
-Using a kitchen stove
21/25

See, the issue is not general cognitive decline (at least in this study), but disruption to a highly specialized skill vastly important for very particular tasks.

That's an important distinction.
22/25

This suggests the need for more research on spatial encoding and other plausible processes (attention, processing speed, spatial skills) & neuroimaging focused on what's happening in areas like DG-mediated mossy fibers in the CA3 region of the hippocampus.
23/25

Take-home points:
-Declines in spatial encoding, not general cognition
-Be careful to avoid claims about "IQ"
-Need more research to rule out more general cognitive processes
-Need neuroimaging studies
24/25

Link to the article:

25/25
thelancet.com/journals/eclin…