The Collision of Cognition : A Revolution

The cognitive revolution did not begin in occupational therapy. It began in a moment of collective scientific crisis, when researchers across psychology, linguistics, computer science, and neuroscience simultaneously recognized that describing behavior was not the same as explaining it. George Miller captured this precisely when he identified September 11, 1956, at MIT as the moment cognitive science was conceived. It was a single afternoon when Chomsky dismantled behaviorist accounts of language, Newell and Simon demonstrated machine problem-solving, and Miller himself presented the limits of human information processing. No one in that room called it a revolution while it was happening. But the underlying premise was radical: the human mind was not only a legitimate object of scientific inquiry, it was the central one. What followed was not a single discipline advancing. It was six fields colliding, each having independently arrived at the conclusion that the solution to their most important problems lived in someone else's territory. That collision produced cognitive science. Occupational therapy did not produce it. But Claudia Allen understood it, entered it, and asked the one question none of those fields had thought to ask: what does cognition look like when a person has to actually do something?

That question was not soft. It was not a retreat into occupational comfort. It was a precision instrument pointed directly at the gap cognitive science had left open. The behaviorists had studied observable action without internal process. The cognitivists had studied internal process in controlled laboratory conditions stripped of real-world demand. Neither camp was asking what happened when those two things collided in a person trying to make breakfast, manage medication, or sequence a morning routine. Allen's insight was that occupation was not a soft alternative to rigorous cognitive science. It was the stress test cognitive science had not yet run. The leather-lacing task was not craft. It was a performance-based window into how a person attends to sensory cues, processes them, and executes voluntary motor action, precisely the input-throughput-output model the Allen Cognitive Disabilities Model articulated formally. She was not borrowing loosely from psychology. She was operationalizing information processing theory through the lens that only occupational therapy possessed: the lens of real functional demand.

The problem is what happened next. The field received Allen's framework as an occupational therapy idea rather than as occupational therapy's contribution to a broader scientific conversation. And once we housed it entirely inside our own discipline, we stopped asking the harder question: does this framework hold up against the standards the rest of cognitive science applies to itself?

Instructive and Uncomfortable

The research record that has accumulated since then is instructive and uncomfortable in equal measure. The scoping review by Liu and colleagues, synthesizing 24 studies and 1,951 participants across six diagnostic groups, found that of 117 quantitative relationships examined between ACLS scores and standardized cognitive assessments, 73.5% were statistically significant. The strongest associations were with global cognition, social cognition, executive function, and attention (i.e., domains that map directly onto integrated, purposeful cognitive processing). The weakest associations were with memory encoding, visuospatial perception, and orientation (i.e., domains that assess discrete cognitive capacities in decontextualized conditions). That pattern is not a validity problem. It is construct evidence. The ACLS appears to be measuring something real: observable sensorimotor information processing during purposeful activity, which is distinct from but systematically related to the internal cognitive capacities traditional neuropsychology quantifies. That is a defensible scientific position. But we have rarely defended it on those terms, because doing so requires us to sit in the neuropsychology literature, the psychometrics literature, and the cognitive science literature long enough to argue from it rather than alongside it. 

The Wesson validity study makes this tension explicit. In a community sample of 160 older adults spanning cognitively normal, MCI, and mild dementia, the LACLS-5 demonstrated strong discriminative validity between dementia and both other groups, with AUCs ranging from 0.80 to 0.89. It was significantly associated with AMPS Process scores, Block Design, Logical Memory, Trail Making Test B, and MMSE. This is a pattern of convergent validity that tells us the instrument is capturing something meaningfully related to visuospatial integration, executive processing, and global cognitive status. But it failed to reliably distinguish between cognitively normal and MCI populations, with an AUC of only 0.63. That is not a failure to sweep under the rug with a clinical anecdote. That is a finding that demands we ask why the instrument's sensitivity breaks down at that boundary, what cognitive demands would need to be introduced to detect subtler impairment, whether scoring refinements or timed components would extend its discriminative reach, and whether the ACLS-6 introduces task complementarity that addresses this gap (which was also a question the scoping review flags urgently, given that only one of the 24 included studies examined the current ACLS-6 version).

These are the questions a science-generating profession asks. They require us to know psychometrics, to understand item response theory, to engage with the COSMIN taxonomy of construct validity, and to read the neuropsychological literature on MCI detection with enough fluency to identify what we can contribute that no other assessment approach can. We have not consistently done that. We have more often described the clinical utility of Allen's levels and moved on.

The original David and Riley study from 1990 pointed toward what rigorous engagement looks like. They correlated ACL scores with the Symbol-Digit Modalities Test, the Shipley Institute of Living Scale, the Beck Depression Inventory, and the MMPI in a psychiatric sample and found that the ACL's strongest association was with visual-motor speed and concentration rather than abstract reasoning or verbal skills which suggests the instrument measures perceptual-integrative, sensorimotor cognitive functioning rather than crystallized intellectual capacity. That finding has direct implications for how we interpret ACL scores, how we communicate them to multidisciplinary teams, and how we position the ACLS in a comprehensive cognitive evaluation. It also raises a question that the 2026 scoping review returned to thirty-six years later: if the ACLS measures integrated, sensorimotor information processing during purposeful activity, and traditional neuropsychology measures discrete internal cognitive capacities, then these two approaches are not in competition, they are measuring different aspects of the same person through epistemologically different lenses.

Functional cognition, as the scoping review's authors conclude, appears to be a construct distinct from both fluid and crystallized cognition. That is a meaningful scientific claim. But it only carries weight if we build the structural validity evidence to support it, if we examine whether the ACLS's hierarchical scoring structure empirically corresponds to the theoretical cognitive levels it proposes, and if we generate the longitudinal data that would tell us whether ACLS scores predict real-world functional outcomes over time. That means not just cross-sectionally, but as people's cognition changes.

None of that work can be done from inside occupational therapy alone. This is precisely what the podcast's central argument demands we reckon with: Allen did not produce the Cognitive Disabilities Model by staying within the boundaries of OT literature. She produced it by entering the scientific conversation of her era with enough intellectual preparation to ask a question that conversation had not thought to ask. The revolution she responded to was built by people who were willing to violate disciplinary boundaries in service of a better explanation. Miller was a psychologist who became a cognitive scientist. Chomsky was a linguist who rewrote how we understood mental structure. Newell and Simon were computer scientists who built models of human problem solving. None of them waited for consensus to form before they contributed. They contributed because they had done enough work in adjacent fields to recognize where the gap was.

The gap right now is not small.

Precision rehabilitation, network neuroscience, computational modeling, digital biomarkers, and AI-assisted assessment are redefining what clinical cognitive measurement can look like. The field is actively in the process of asking which constructs matter, which measurement approaches are valid, and what the relationship is between performance-based functional assessment and underlying neural architecture.

Occupational therapy has something to contribute to that conversation that no other field can: we understand what cognition looks like when a person has to sustain it across the demands of a real day, in a real environment, with real adaptive requirements. We understand activity analysis at a level of precision that basic science researchers do not have access to. We understand the difference between what a person can do in a structured clinic condition and what that performance predicts about how they will function when they go home. Those are not soft clinical observations. They are measurement science questions. But we will only be taken seriously as contributors to those questions if we arrive at the table with the same rigor the other disciplines expect of themselves, which means publishing in journals outside of OT, building collaborative research programs with neuroscientists and psychometricians and computational modelers, and treating the gaps in our own evidence base not as someone else's problem to solve but as our scientific responsibility to address.

A Seat At The Table, not a Late Arrival

The history is clear on what happens when we do not. The cognitive revolution happened. OT responded to it magnificently, a decade into the conversation. The evidence-based practice movement happened. OT adapted to it, largely on medicine's terms. The neuroscience revolution is happening now. We can wait again, let the frameworks solidify, and translate them into occupation-based applications once the scientific community has moved on to the next question. Or we can take seriously what Allen actually modeled: not the safety of staying in our lane, but the courage of entering a conversation that was already happening and contributing something the rest of the field did not know it needed. Functional cognition is a real construct. The evidence is beginning to support it. But evidence that supports a construct is not the same as a research program that builds one. If occupational therapy is going to lead in this space, we have to do the structural validity work, the longitudinal outcome work, the computational integration work, and the interdisciplinary publication work that transforms a clinically compelling idea into a scientifically defensible framework. That is not a threat to occupation. It is the only way occupation stays at the center of the conversation rather than arriving after it ends.

Key References: 

David, S. K., & Riley, W. T. (1990). The Relationship of the Allen Cognitive Level Test to Cognitive Abilities and Psychopathology. The American journal of occupational therapy, 44(6), 493–497. https://doi.org/10.5014/ajot.44.6.493

Miller, G. A. (2003). The cognitive revolution: a historical perspective. Trends Cogn Sci, 7(3), 141–144. https://doi.org/10.1016/s1364-6613(03)00029-9

Wesson, J., Clemson, L., Crawford, J. D., Kochan, N. A., Brodaty, H., & Reppermund, S. (2017). Measurement of Functional Cognition and Complex Everyday Activities in Older Adults with Mild Cognitive Impairment and Mild Dementia: Validity of the Large Allen's Cognitive Level Screen. Am J Geriatr Psychiatry, 25(5), 471–482. https://doi.org/10.1016/j.jagp.2016.11.021

Liu, G. H., Pordage, L., Pan, T. C., Chou, L. W., & Ma, W. F. (2026). Mapping the Cognitive Constructs Assessed by the Allen Cognitive Level Screen: A Scoping Review. Psychol Res Behav Manag, 19, 593468. https://doi.org/10.2147/prbm.S593468