Neuropsychology
Six fictional language profiles. Put any two side by side and work out which single measure separates them - and then work out what a separation is worth.
Simulated — invented profiles, original task descriptions, no clinical materials
By the end you should be able to say which measure carries most information about a particular contrast, and what a separation is worth.
About 30 minutes. Nothing you do here is saved or sent anywhere.
The tool never names a syndrome, no standardised battery, subtest or item is reproduced, and every task description is original. The six profiles are invented and nothing here assesses anybody.
Two fictional people both speak in short, effortful phrases and both understand what is said to them. One can repeat a spoken sentence back accurately and the other cannot.
| Measure | What it asks for | First | Second | Separation |
|---|
Scores run from 0 to 10, where 10 is performance indistinguishable from a fictional control group. They are illustrative values invented for teaching, not measurements of anything.
Profile F is moderately fluent, moderately impaired at understanding, moderately impaired at repetition, poor at naming, good at reading aloud and very poor at writing. It matches no textbook pattern.
Repetition separates profiles A and D almost perfectly and tells you nothing at all about A against B. Comprehension does the reverse. There is no such thing as the most informative language measure. There is only the most informative measure for the question you are asking, which is why hypothesis-driven assessment beats running a long battery and looking at what comes out.
Repeating a spoken word can be done by understanding it and saying it, or by a shorter route from the sounds coming in to the sounds going out, without meaning being involved. A person whose repetition is disproportionately poor while comprehension and speech are relatively spared is showing that the short route is unavailable - and that is a claim about a connection rather than about a region. A person whose repetition is disproportionately good is showing the opposite: the short route is intact and doing the work while the rest is not.
The classical categories were defined by which symptoms tended to occur together in the cases available a century ago. They are useful shorthand with three well-known problems. A substantial minority of people fit none of them, the same label covers very different profiles, and the anatomical claims attached to them are much looser than the diagrams suggest. Profile F is not an anomaly; it is what a continuous space looks like when you try to sort it into boxes.
"Fluent" bundles rate of speech, phrase length, effort, articulation, grammatical structure and how much content the words carry. Two people rated equally fluent can differ on every one of those, and a single fluency score hides it. The same is true of every measure here - each one is a compound, and a difference on it is a difference in the compound.
The tool mentions no anatomy anywhere. Language depends on a distributed left-hemisphere network and on the white matter that joins it. The same behavioural profile follows damage in several places, and the classical region-to-syndrome mapping has not survived the imaging evidence in the form it was taught. A profile constrains what has gone wrong with the processing. It does not constrain the anatomy nearly as tightly as the textbook diagrams imply.