A Field Guide to My Own Attention

My vault backs itself up whenever I stop working long enough for it to notice. The log is a small dataset: 01:33. 02:39. 07:25. Exit stamps, the moments I surfaced from a stretch of work and the machine recorded how long I'd been gone.

I've had that kind of attention my whole life. Six hours vanish into a type scale; a research question opens on a Tuesday and eats the week. The same nervous system registers the refrigerator's hum from two rooms away and files it as signal, a supermarket at rush hour costs more than the groceries, and a client-facing week carries a predictable recovery cost of about a day.

For most of my life the file name for all of this was "too much." Then I started reading the research and kept meeting myself in papers: sentences written decades ago describing, with clinical precision, patterns I had spent twenty years treating as character flaws. The current shorthand for what I am is AuADHD, an autistic mind and an ADHD mind in one head. This essay holds my inner reports up against the literature that names them, and it follows the rule from the essay I wrote about introspection: the intuition is the cheap part.

The thesis

Stated plainly, since the rest of the essay has to earn it:

• The traits above are not one temperament. They decompose into distinct constructs: sensory over-responsivity, atypical interoception, monotropic attention, autistic burnout, each with its own literature and its own evidential status.
• The framework that unifies the most of my data with the fewest assumptions is monotropism: autism understood as a strategy for distributing scarce attention. It spent two decades as community-endorsed theory; since 2023 it has a validated psychometric instrument.
• Where autism and ADHD overlap, the science is barely a decade old, because the diagnostic manual forbade the dual diagnosis until 2013. First-person reports currently carry information the studies don't yet hold, and I'll use them, labeled.
• None of this supports a superpower narrative. The honest conclusions are methodological (the research over-samples people like me) and practical (you don't fix an architecture; you engineer around it).

Method, in one line: well-established findings cited plainly; contested findings flagged in the same sentence; lived experience marked as exactly that, hypothesis-generating, n = 1.

Sensory processing: the firmest ground

Sensory differences sit in the diagnostic criteria themselves, and the mechanistic work treats them as central rather than incidental. Two frameworks are load-bearing.

The first is excitation/inhibition balance: the hypothesis that autistic cortex runs at a higher ratio of excitatory to inhibitory signaling, less damping, more gain. It's a framework for organizing findings, not a settled result, and it's still being tested at the circuit level. Subjectively it describes my situation exactly: the gain is set high and is not voluntarily adjustable.

The second is Bayesian. Perception is inference: prior expectation plus incoming evidence, weighted. Pellicano and Burr propose that autistic perception runs on attenuated priors, "hypo-priors": less smoothing toward the expected, more weight on the raw signal, and their phrase for the consequence is that the world becomes "too real". That is the most precise three-word description of a supermarket I have encountered in print.

There is a third account, the Intense World Theory: hyper-reactive, hyper-plastic local circuits producing a mind that experiences too much and withdraws to cope. It is the account that flatters my self-description most and the one I trust least, for the same reason. It was built outward from a valproic-acid rodent model, and it remains influential, criticized, and unconfirmed. I file it as hypothesis, not result.

Interoception: where the evidence genuinely conflicts

The strongest claim I'd like to make is that I'm unusually tuned into my own internal states. The literature does not permit it, and the way it refuses is worth walking through, because it's a model case of why self-report needs an audit.

Interoception fractionates. Garfinkel and colleagues split it into three dimensions that correlate surprisingly poorly: accuracy (objective performance, classically a heartbeat-counting task), sensibility (self-reported attunement), and awareness (the metacognitive calibration between the two). Autism studies disagree partly because they measure different dimensions: several find reduced accuracy, others elevated sensibility. And one confound may dissolve the autism effect entirely: Shah, Bird and colleagues found that impaired interoceptive accuracy tracks alexithymia, not autism, a co-occurring difficulty identifying emotion that many autistic people have and many don't. Control for alexithymia and the autism difference shrinks toward zero.

So the defensible version of my claim is: high interoceptive sensibility, unverified accuracy. A loud signal is not the same as a calibrated one. In the language of the earlier essay: this hypothesis is still in review.

Monotropism: the model that fits

In 2005, Dinah Murray, Mike Lesser and Wenn Lawson proposed that autism's diagnostic features fall out of one variable: how attention is distributed across a person's interest system. Attention is scarce. A polytropic strategy spreads it across many interests at low arousal; a monotropic strategy concentrates it into few interests at high arousal. From that single difference the model derives the clinical picture: deep flow inside the attention tunnel, high task-switching costs, distress at interruption, slow re-entry, and "restricted interests" as the outside view of a resource-allocation strategy.

As an explanation of my data, it is embarrassingly efficient. The flow states that swallow nights: predicted. The exit stamps: predicted. Interruption experienced as decompression rather than pause: predicted. Even the vault, a few-thousand-note research graph running from art history through consciousness studies to physics, stops looking like scatter and starts looking like what a monotropic system does when its interest is systems themselves.

For two decades this was the theory the autistic community endorsed while academia mostly looked elsewhere. The psychometrics have begun to arrive: in 2023, the Monotropism Questionnaire was validated on 1,110 participants (756 autistic), a 47-item instrument with an eight-factor structure, high internal consistency, and strong convergent validity with established autism measures (correlations above .61). Community recognition was the first kind of validity; this is the second kind arriving.

Reading the 2005 paper felt less like discovery than like correction, twenty years late.

AuADHD: young literature, real overlap

Until DSM-5 in 2013, autism was an exclusion criterion for ADHD: you could officially be one or the other, not both. That editorial decision is why the co-occurrence literature is barely a decade old. What it shows so far: an estimated 30 to 50 percent of autistic individuals show clinically relevant ADHD symptoms, a substantial share of ADHD-ers show autistic features, and twin and family studies put the genetic overlap between the two conditions at roughly 50 to 72 percent. Common, heritable, and structurally under-studied.

What the combination is like from the inside is currently better documented by the community than by the journals, so mark this paragraph experiential. The two attention systems do not cancel out; they interact. The autistic system prefers one deep, long channel and pays heavily for severance. The ADHD system gates on interest and salience: it does not sustain on obligation, and it starves on repetition. Aligned on one target, they produce hyperfocus of unusual depth. Misaligned, they produce the characteristic AuADHD weather: under-stimulated and over-stimulated in the same hour.

One early empirical anchor suggests the overlap may run through attention itself: in the Monotropism Questionnaire validation, autism and ADHD status were each independently associated with higher monotropism scores. The single-channel pattern may be the thing the two conditions share, reached from different directions.

Read through this lens, my project constellation, an agency, a studio, a research wiki, mathematics I'm relearning, symbolic systems studied seriously, is not distractibility. It's serial monotropy: many tunnels, entered one at a time, selected by interest.

Cognitive style: detail-first

Three models describe the local bias in autistic cognition. Mottron's Enhanced Perceptual Functioning treats it strength-first: perception operates with more autonomy from top-down control and outperforms on tasks where detail wins, visual search, embedded figures, pitch discrimination. Weak central coherence began as the deficit version, "sees trees, misses forest," and was revised by its own authors into a style claim: a default toward local processing, with global processing available on demand. Both are partially supported and actively debated, and I cite them as such. (Empathizing–systemizing theory also exists; it arrives wrapped in the "extreme male brain" framing and its substantial criticisms01904-6), so I take the one useful construct, systemizing, and leave the rest.)

The practical version in my work: structure arrives before surface. I recently wrote an entire essay about designing the bones of a document before its face, as professional method, without noticing it was also a cognitive self-portrait. Design turns out to be a profession whose job description matches a detail-first perceptual style unusually well.

Reasoning: two kinds of intuition

There's an apparent contradiction to clear before this map is honest. The reasoning literature reports that autistic people are less intuitive, and I reason for a living on something I'd call intuition. Both are true, because the word is doing two different jobs.

Brosnan and colleagues gave autistic and non-autistic adults the Cognitive Reflection Test and found the autistic group answered less intuitively and more deliberatively, at medium-to-large effect sizes. Two facts about that finding matter more than the finding itself. On the CRT the "intuitive" answer is the wrong answer, the lazy heuristic snap the test is built to bait, what the field calls cognitive miserliness; "less intuitive" means less prone to the shortcut that produces the error. And the study is small, seventeen versus eighteen males, with no control for cognitive ability, a limitation its own authors flagged. When a large pre-registered study added that control, Taylor and colleagues, roughly 1,200 participants, watched the objective difference disappear; the only durable result was that autistic people report being less intuitive. Like the interoception evidence, the flattering version doesn't survive a proper control. File it there.

The distinction that does survive is the one worth having. Dual-process intuition, the CRT kind, is the bias-prone snap judgment. Expertise-based intuition is a different construct: the fast pattern-recognition an expert builds through long immersion in a domain that gives real feedback, what Herbert Simon called analyses frozen into habit. It is deliberative knowledge that became fast by being run ten thousand times. From opposite camps, Kahneman and Klein agreed in 2009 that this intuition is real and conditional: it develops only where the environment is regular enough to learn and the feedback honest enough to correct you, the same deliberate-practice engine Ericsson described, though how much of expertise that engine explains is itself contested.

So the two claims reconcile without strain. I am, plausibly, a reasoner who resists the lazy heuristic; I don't experience myself as a gut-feeling thinker, which fits the one finding that replicates. And I lean hard on the earned intuition, the recognition that arrives fast because it was assembled slowly. Monotropic immersion is exactly the raw material that kind is made from: sustained, narrow, high-feedback attention, run for years, on interest instead of obligation. No study has walked the path from monotropism to expert recognition, so I state it as synthesis, not citation, but it is the cleanest account I have of why the deep tunnel produces judgment I can mostly trust.

This is also where the earlier essay's rule takes its correction. "The intuition is the cheap part" is true of the naive kind, the one I'm apparently short on. The earned kind is the expensive part, and there is no shortcut to it. The coda I've made my peace with is Kahneman and Klein's: there is no subjective marker, no inner signal that certifies a confident intuition as earned rather than biased. From the inside, the two feel identical. Only the regularity of the domain and the quality of the feedback can tell you which one you are running, which is, more or less, the entire argument for testing what your mind hands you instead of trusting it.

The cost function

Every parameter above has a cost term, and omitting the cost terms is how this genre degrades into inspiration content.

Overload is arithmetic, not drama: higher gain on input, finite processing capacity, and past threshold the system degrades, with shutdown functioning as a circuit breaker. Masking, running social interaction as explicit computation instead of background process, draws on the same finite budget; it is the unpaid second job that funds the crash. And the long-run failure mode now has a defined construct: Raymaker and colleagues, working with community-based participatory methods (thematic analysis of 19 interviews plus 19 public first-person accounts), define autistic burnout as chronic exhaustion, skill loss, and reduced stimulus tolerance persisting three months or more, produced by chronic life stress and a mismatch between expectations and capacity without adequate support. The paper's title is a participant quote: "having all of your internal resources exhausted beyond measure and being left with no clean-up crew."

My planning treats the recovery day after a client-heavy week as a fixed cost. Amortized, it's cheaper than the accumulated alternative.

Why none of this is a superpower argument

Two reasons, one ethical, one methodological, and the methodological one cuts deeper.

The ethical critique comes from inside the autistic community: superpower framing quietly deletes autistic people with high support needs and makes acceptance conditional on economic usefulness. The community's own long-run position, difference and disability held simultaneously, has empirical backing in how autistic adults actually describe themselves.

The methodological point: autism research systematically over-samples verbal, testable, average-or-higher-IQ participants. Russell and colleagues quantified this in a cross-sectional review and meta-analysis of selection bias: study samples skew heavily toward the end of the spectrum that can sit through studies. Which means the published picture of autism, articulate, pattern-loving, employable, is partly an artifact of recruitment. When I recognize myself in the literature, sampling bias predicts that I would. The privilege caveat is a correction term, not etiquette.

The functioning label doesn't survive contact with data either. In a sample of 2,225 autistic children, IQ was a poor predictor of adaptive functioning: the group without intellectual disability showed everyday-life skills far below what their cognitive scores implied, and the authors recommend retiring "high-functioning" outright. I function well in a life engineered to my own specification, and I had the resources to do the engineering. Remove that clause and the same nervous system yields a very different outcome.

Engineering, not narrative

The practical output of a map like this is a set of design constraints. Mine, currently: long uninterrupted blocks, ideally one channel per day, because task-switching is the single most expensive operation in the system. Controlled sensory input. Externalized state: the vault holds the open loops so working memory doesn't have to, and contexts get stored rather than switched, then re-entered later.

The newest constraint is the one I'd defend hardest, and it has no literature behind it, so weight it accordingly: I use AI as a polytropic prosthesis. The failure mode of a monotropic system running a studio is not the deep work; it's the wide work, the many shallow concurrent threads of follow-ups, publishing steps, status tracking, administrative upkeep. Shallow breadth is exactly what a language model supplies cheaply. So the division of labor follows the architecture: the model holds the wide, I hold the deep, and re-entry after a tunnel costs one question instead of an hour of reconstruction. One caution rides along with it, the one the reasoning above implies: expertise-based recognition stays trustworthy only while the feedback stays honest, and a tool that drafts what I would have drafted can quietly remove the feedback I'd otherwise have learned from. So the standing rule is to verify the model's output, not to trust the recognition it might be eroding. Explicitly n = 1, possibly a crutch with long-term costs I can't see yet. What I can report is that the architecture I was told to fix now has infrastructure instead, and the difference is not subtle.

That is the whole practical thesis: you don't argue with an architecture. You read its spec sheet, and you build for it.

Where this leaves me

For years I graded my attention on a curve built for a different architecture and logged the differences as personal failure. The literature, contested exactly where I've flagged it, retired that grading scheme: the same data, re-read under a better model. "Too much" was a measurement error.

The references below are doing load-bearing work. The strongest thing I can say for this field is that it argues with itself, which is what a live science does. The strongest thing I can say for the map is that engineering to it works. Both claims are testable. One of them, I test daily.

References

• Alvares, G. A., Bebbington, K., Cleary, D., Evans, K., Glasson, E. J., Maybery, M. T., Pillar, S., Uljarević, M., Varcin, K., Wray, J., & Whitehouse, A. J. O. (2020). "The misnomer of 'high functioning autism': Intelligence is an imprecise predictor of functional abilities at diagnosis." Autism, 24(1), 221–232. https://journals.sagepub.com/doi/10.1177/1362361319852831
• American Psychiatric Association (2013). Diagnostic and Statistical Manual of Mental Disorders, 5th ed. The first edition to permit a dual autism + ADHD diagnosis.
• Baron-Cohen, S. (2002). "The extreme male brain theory of autism." Trends in Cognitive Sciences, 6(6), 248–254. https://doi.org/10.1016/S1364-6613(02)01904-6 (Cited here as contested; see the substantial criticisms of the empathizing–systemizing framing.)
• Brosnan, M., Lewton, M., & Ashwin, C. (2016). "Reasoning on the autism spectrum: A dual process theory account." Journal of Autism and Developmental Disorders, 46(6), 2115–2125. https://doi.org/10.1007/s10803-016-2742-4 (Study 2: n = 17 vs 18 males, no IQ control; the CRT "intuitive" response is, by definition, the error.)
• Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). "The role of deliberate practice in the acquisition of expert performance." Psychological Review, 100(3), 363–406. https://doi.org/10.1037/0033-295X.100.3.363
• Garau, V., Murray, A. L., Woods, R., Chown, N., Hallett, S., Murray, F., Wood, R., & Fletcher-Watson, S. (2023). "Development and Validation of a Novel Self-Report Measure of Monotropism in Autistic and Non-Autistic People: The Monotropism Questionnaire." OSF preprint. https://osf.io/ft73y/ · Questionnaire: https://monotropism.org/questionnaire/
• Garfinkel, S. N., Seth, A. K., Barrett, A. B., Suzuki, K., & Critchley, H. D. (2015). "Knowing your own heart: Distinguishing interoceptive accuracy from interoceptive awareness." Biological Psychology, 104, 65–74. https://doi.org/10.1016/j.biopsycho.2014.11.004
• Grandin, T. (2006). Thinking in Pictures, expanded ed. Vintage. First-person account of visual/associative processing.
• Happé, F., & Frith, U. (2006). "The weak coherence account: Detail-focused cognitive style in autism spectrum disorders." Journal of Autism and Developmental Disorders, 36(1), 5–25. https://doi.org/10.1007/s10803-005-0039-0 (Presented as debated, per its own authors' reframing.)
• Henrich, J., Heine, S. J., & Norenzayan, A. (2010). "The weirdest people in the world?" Behavioral and Brain Sciences, 33(2–3), 61–83. https://doi.org/10.1017/S0140525X0999152X
• Kahneman, D., & Klein, G. (2009). "Conditions for intuitive expertise: A failure to disagree." American Psychologist, 64(6), 515–526. https://doi.org/10.1037/a0016755 (The adversarial collaboration that set the boundary conditions for trustworthy expert intuition.)
• Kapp, S. K., Gillespie-Lynch, K., Sherman, L. E., & Hutman, T. (2013). "Deficit, difference, or both? Autism and neurodiversity." Developmental Psychology, 49(1), 59–71. https://doi.org/10.1037/a0028353
• Lawson, W. (2011). The Passionate Mind: How People with Autism Learn. Jessica Kingsley. Monotropism from lived experience.
• Leitner, Y. (2014). "The co-occurrence of autism and attention deficit hyperactivity disorder in children – what do we know?" Frontiers in Human Neuroscience, 8, 268. https://www.frontiersin.org/journals/human-neuroscience/articles/10.3389/fnhum.2014.00268/full
• Macnamara, B. N., & Maitra, M. (2019). "The role of deliberate practice in expert performance: Revisiting Ericsson, Krampe, and Tesch-Römer (1993)." Royal Society Open Science, 6(8), 190327. https://doi.org/10.1098/rsos.190327 (Reanalysis; deliberate practice explains less performance variance than originally claimed.)
• Markram, K., & Markram, H. (2010). "The Intense World Theory – a unifying theory of the neurobiology of autism." Frontiers in Human Neuroscience, 4, 224. https://doi.org/10.3389/fnhum.2010.00224 (A hypothesis: influential, widely criticized, not consensus.)
• Mottron, L., Dawson, M., Soulières, I., Hubert, B., & Burack, J. (2006). "Enhanced perceptual functioning in autism: An update, and eight principles of autistic perception." Journal of Autism and Developmental Disorders, 36(1), 27–43. https://doi.org/10.1007/s10803-005-0040-7
• Murray, D., Lesser, M., & Lawson, W. (2005). "Attention, monotropism and the diagnostic criteria for autism." Autism, 9(2), 139–156. https://doi.org/10.1177/1362361305051398
• Murray, F. (2019). "Me and Monotropism: A unified theory of autism." The Psychologist, 32, 44–49. https://www.bps.org.uk/psychologist/me-and-monotropism-unified-theory-autism · Community resources: https://monotropism.org/
• Pellicano, E., & Burr, D. (2012). "When the world becomes 'too real': A Bayesian explanation of autistic perception." Trends in Cognitive Sciences, 16(10), 504–510. https://doi.org/10.1016/j.tics.2012.08.009
• Raymaker, D. M., Teo, A. R., Steckler, N. A., Lentz, B., Scharer, M., Delos Santos, A., Kapp, S. K., Hunter, M., Joyce, A., & Nicolaidis, C. (2020). "'Having All of Your Internal Resources Exhausted Beyond Measure and Being Left with No Clean-Up Crew': Defining Autistic Burnout." Autism in Adulthood, 2(2), 132–143. https://www.liebertpub.com/doi/full/10.1089/aut.2019.0079
• Robertson, C. E., & Baron-Cohen, S. (2017). "Sensory perception in autism." Nature Reviews Neuroscience, 18(11), 671–684. https://doi.org/10.1038/nrn.2017.112
• Russell, G., Mandy, W., Elliott, D., White, R., Pittwood, T., & Ford, T. (2019). "Selection bias on intellectual ability in autism research: A cross-sectional review and meta-analysis." Molecular Autism, 10, 9. https://doi.org/10.1186/s13229-019-0260-x
• Sohal, V. S., & Rubenstein, J. L. R. (2019). "Excitation-inhibition balance as a framework for investigating mechanisms in neuropsychiatric disorders." Molecular Psychiatry, 24(9), 1248–1257. https://doi.org/10.1038/s41380-019-0426-0
• Taylor, E. C., Farmer, G. D., Livingston, L. A., Callan, M. J., & Shah, P. (2022). "Rethinking fast and slow processing in autism." Journal of Psychopathology and Clinical Science, 131(4), 392–406. https://doi.org/10.1037/abn0000734 (Pre-registered, cognitive-ability-controlled; no objective autism difference beyond lower self-reported intuition.)