Bipolar Disorder

Case: Simone, 16. Five chapters follow: the case, epidemiology, presentation, functional neuroscience, and treatment.

The case: Simone, 16

Presenting concern

Simone is sixteen and in eleventh grade. Her pediatrician wrote: "I started fluoxetine in July for recurrent depression and raised it to 40 mg in late August. For twelve days Simone has slept three to four hours a night, talks faster than her family can follow, and has painted her bedroom walls. Is it the medication? I've held it until you see her."

Her mother: "She's happy. That's what frightens me. She hasn't been happy in three years, and I don't recognize it."

Simone, seen alone: "I finally have the energy I was supposed to have. Everyone says slow down; why would I?" About sleep: "It's a waste. I lie down and my head is already drawing the next page." She denied wanting to die: "For the first time in years, I don't."

Developmental and family history

In sixth grade a teacher called her "bright, daydreamy, never quite with us"; no evaluation followed. At thirteen, after a move, three months of withdrawal, long sleep, and no drawing resolved untreated; her mother dates a change from then: "She never fully came back. She's been gray since." In tenth grade a clearer depressive episode brought low mood, twelve hours of sleep, weight gain, and thoughts that she would rather not be alive, without plan; five months of cognitive behavioral therapy brought partial recovery. Low mood returned in July; fluoxetine was started, then raised in late August because she was "still flat." Her mother recalls a three-day weekend at fourteen when Simone rearranged every room in the house and "talked our ears off," followed by an ordinary week. No other period of sustained elevated mood is identified.

Her mother's sister has bipolar I disorder, was hospitalized twice in her twenties, and has been well on lithium for years. Her maternal grandfather "had big ideas and drank" and lost a business twice. Her brother, Ollie, is thirteen.

Simone drinks two or three energy drinks a day, a habit begun during May finals, when she also twice took a friend's prescribed stimulant. She denies any since and vapes nicotine; a urine drug screen is pending.

Observations in session

Simone arrived with forty inked pages of a graphic novel begun eleven days ago, spoke quickly, and was hard to interrupt, though she answered each question eventually. She moved from the book to selling prints to fund "a studio" to an email she sent an editor at 3:40 a.m. ("They'll answer. It's that good."). Asked whether she might be overestimating, she laughed: "Maybe. I don't think so." Her mood was "the best it's been," her thoughts "faster than I can draw them." She denied hallucinations and thought insertion. "Sometimes I know a page is going to matter before I draw it"; pressed, she described certainty, not a message from anywhere. She has spent $340 of savings on paint and has messaged an illustrator she admires eleven times in two days.

Asked to put the sketchbook away, her face changed. "You're doing what my mother does," she said, and for about a minute her tone was sharp. Then she brightened. She could not stay on one topic, yet recounted her project's sequence accurately and in order.

Collateral

Her mother's sleep log this week: in bed between three and four, up at seven, "not tired, bouncing down the stairs." During finals last May, by contrast, Simone stayed up until one or two for five nights on coffee and energy drinks, "was a wreck," and slept fourteen hours the following Saturday.

She also described two kinds of anger. Simone and Ollie "have fought every day for years" about the bathroom, the remote, and noise; the fights are loud, short, predictable, and "that's just them." Last Thursday, asked to stop painting and come to dinner, Simone screamed, threw a brush at the wall, and said, "You're stealing this from me." Ten minutes later she came downstairs cheerful, planning a dinner for everyone.

Her art teacher wrote that Simone "has produced more in two weeks than in all of last year," interrupts constantly, and has started three murals in the art room without asking, none finished. Her therapist from last year said Simone was irritable with her brother during the depression too, "but then she hid from him. Now she's chasing him around with the sketchbook."

Epidemiology

How common

The NCS-A found a lifetime prevalence of bipolar I or II disorder of 2.9% among 10,123 US adolescents aged 13 to 18 by structured interview [53]. That is the only community prevalence figure in the reference list, and it is a lifetime figure that includes hypomania. Simone is 16, and the figure says that in a school of a thousand students, on the order of thirty will have met criteria by the time they are 13 to 18; it does not say how many of them had the episodic, distinct-period illness that the presentation chapter describes, because the survey's figure combines bipolar I and II and does not separate the phenotypes.

The diagnosis in clinics

A 2008 review reports that the rate of the diagnosis doubled in outpatient settings over the preceding decade and that twice as many research articles on pediatric bipolar disorder appeared in the five years before the review as in the ten years before that [66]. The review concentrates on youths with severe chronic irritability, hyperarousal and hyperreactivity, the group in whom the diagnosis was and remains most debated [66]. The rise in clinical diagnoses and the community figure are not in conflict: one counts what clinicians write, the other what a standardized interview finds, and the gap between them is the epidemiological expression of the phenotype question.

The follow-up data cut against diagnosing chronic irritability as bipolar disorder. In adult follow-up reported in a DMDD review, children with the chronic-irritability presentation had elevated rates of anxiety and depression as adults relative to psychiatric comparisons, and bipolar disorder was not among the elevated outcomes [7]. The DMDD chapter carries the full figures.

Who develops it

Two retrieved sources describe antecedents. In a prospective study of 55 clinically referred children whose first depression was dysthymic disorder, 13% went on to develop bipolar disorder over 3 to 12 years of follow-up, alongside 76% who developed first-episode major depression [18]. In the Dunedin cohort, adult disorders were preceded by juvenile diagnoses in 73.9% of cases, frequently of a different kind [59]; the cohort's abstract does not report bipolar disorder separately. Simone's history, with two prior depressive episodes, a low-grade course since thirteen and an affected maternal aunt, is the profile these sources describe, and none of them provides a probability for her.

Psychotic features

Psychotic features are common in youth bipolar disorder and vary widely by sample: a review cites a range of 16% to 87.5%, higher in adolescent-onset than childhood-onset cases, compared with about 4% in major depressive disorder [89]. Among 215 children referred to the NIMH childhood-onset schizophrenia study with psychosis, 15.2% received a diagnosis of bipolar disorder, alongside 30% schizophrenia, 12% major depression with psychotic features and 4.6% schizoaffective disorder [89]. Psychotic symptoms in an adolescent with an activated episode therefore do not, by themselves, move the diagnosis toward a primary psychotic disorder.

What is missing from this list

The reference list holds no source on the sex distribution, age-of-onset distribution or family-history risk of bipolar disorder in youth, and none on the rate of bipolar outcome after antidepressant-associated activation. Those gaps are left open here rather than filled; the presentation chapter's list of claims needing sources does not yet include them.

What this case cannot establish

A single activated episode following an antidepressant dose increase, in a girl with a family history, sits at the center of a diagnostic question the base rates cannot settle. The prevalence figure makes bipolar disorder plausible; it does not make it likely, and the case chapters are built on what the episode itself shows.

Source status

#53 from the article text; #66 from the publisher's abstract (abstract only); #7 from open-access full text; #18 from the published abstract; #59 from the abstract; #89 from open-access full text.

Figures for this chapter

Presentation

What was observed, what was reported, what was inferred

The pediatrician's letter contains a sleep log, a rate of speech, and painted walls: observations. It also contains a question, "Is it the medication?", which is an inference waiting for evidence. Simone's mother inferred happiness, then inferred that the happiness is not Simone's. Simone reports energy, speed, certainty, and no wish to sleep. The case turns on comparisons, so assessment begins by keeping the reports apart.

An activated state is a coordination of features

Activated mood states require attention to the coordination of features: elevated or irritable mood, accelerated thought and speech, increased goal-directed activity, altered judgment, and reduced need for sleep. The critical comparison is with the young person's own baseline over a sustained interval; ordinary enthusiasm, an isolated late bedtime, or chronic distractibility cannot carry that inference alone [340].

In Simone the features arrived together and have stayed together: three to four hours of sleep without fatigue, speech her family cannot follow, forty pages and three murals, a 3:40 a.m. email, $340 of savings, eleven messages to a stranger, and a mood she calls the best it has been. Any one alone describes a sixteen-year-old with a project. The baseline is her own: "gray since thirteen," an adolescent who last year slept twelve hours and did not draw. The distance from that baseline, the number of functions that moved, and the twelve days they have stayed moved make this an episode.

Two kinds of late night

Reduced need for sleep means sleeping substantially less while experiencing unusually little need for recovery. A diary that records "six hours" misses whether those hours followed anxious wakefulness, voluntary delay, or a broader episode of activation.

In May, Simone stayed up until one or two for five nights on coffee and energy drinks, was "a wreck," and slept fourteen hours the first Saturday she could. Sleep pressure accumulated, was felt, was resisted for a purpose, and was repaid: voluntary restriction with the need for sleep intact. This month she is in bed between three and four, up at seven, "not tired," and calls sleep "a waste." There is no recovery sleep, no stimulant needed to stay awake (the energy drinks predate the episode and have not increased), and no external purpose that accounts for the hours. The log alone could not have made the separation; it took a mother who could describe May. The observation establishes a reduced need for sleep by report, different in kind from finals week; it does not measure circadian phase or sleep pressure, and it does not say why.

Two kinds of anger

Irritability describes a propensity toward anger or frustration, not simply sadness expressed loudly. The outward response, the felt emotion, and the interpretation of the request deserve separate descriptions.

Simone's fights with Ollie are loud, daily, years old, about the bathroom and the remote, and over quickly for both. They were present during the depression, when she withdrew from him, and now, when she pursues him: a standing pattern in one relationship, predictable in trigger, brief, symmetrical, and not confined to any mood state.

Thursday's dinner differs in each respect. The trigger was an interruption of goal-directed activity, not a sibling. The content was that something was being taken from her. The intensity was out of proportion to the request. The resolution was a return to expansiveness within ten minutes, not to neutral, and the same shape recurred in session when the sketchbook was put away. This irritability appears when the current pursuit is stopped from outside and dissolves when the pursuit resumes; it is bound to the state and runs on the state's clock. From across the kitchen the two look alike: Simone shouting. Recording trigger, content, duration, and the mood that follows keeps them apart.

Valence, energy, speed, and judgment, described separately

Depression and activation need not occupy opposite ends of a happiness scale, so valence, energy, thought speed, and inhibition are described separately. Simone's valence is elevated, with brief irritable intervals when redirected. Her energy holds through the night. Her thought speed is high, yet she recounted the order of her project accurately, so speed has not yet disorganized content. Her judgment is altered with consequences: money, a publisher emailed at an hour that tells the publisher something, a stranger messaged eleven times. Her self-appraisal is inflated but not fixed. "Maybe. I don't think so" is an estimate that hears doubt and sets it aside, not a belief maintained against evidence. Her certainty that "a page is going to matter" is, on questioning, conviction rather than communication from outside, and she denied hallucinations and thought insertion. The primers on psychotic symptoms in youth describe such symptoms as occurring across many conditions, including mood disorders, and direct that they be assessed in their temporal relation to the mood state [87, 89]. For now the mood is elevated and the beliefs are its content; a belief that stops yielding to doubt, or an experience located outside herself, would be recorded as a change.

Observation	Can establish	Cannot establish
Three to four hours of sleep, "not tired," twelve nights	Reduced need for sleep by report, distinct from May	Circadian phase; sleep pressure; cause
Brush thrown at dinner, cheerful ten minutes later	State-bound irritability on interruption	That Simone is "an irritable child"; a mechanism
Daily fights with Ollie	A standing pattern across mood states	Anything about the episode
Forty pages in eleven days	Increased goal-directed activity; preserved skill	Quality of the work
Email to an editor at 3:40 a.m.	Altered judgment about timing and probability	A delusion
"Maybe. I don't think so."	Inflated but revisable self-appraisal	That it will stay revisable

The differential, held open

ADHD was raised in sixth grade and never evaluated. It would predict distractibility and impulsivity present across years and settings at a steady level, not arriving over twelve days with reduced need for sleep and elevated mood. Simone's inattention is new in degree and moved with everything else; whether a lifelong pattern also exists cannot be assessed during an episode.

DMDD and chronic irritability predict persistently irritable mood between outbursts, beginning before age ten, and specifically exclude a distinct period of elevated mood and increased activity lasting more than a day [7]. Simone's sibling fights are chronic, but she is not described as angry between them, and the anger of the past twelve days is confined to interruptions of an elevated state. Chronic irritability versus episodic mania is the central debate of the pediatric bipolar literature; Leibenluft and Rich argued that episodicity, with elevated mood or grandiosity, defines the narrow phenotype that longitudinal and family data best support, while chronic irritability without episodes follows a different course [66]. Simone's presentation is episodic, which places her on one side of the debate without settling what the episode is.

A substance-induced state would follow exposure and resolve with clearance. The energy drinks predate the episode; she denies stimulants since May; the urine screen is pending. Caffeine and nicotine are weighed like any exposure, by timing, dose, and persistence against the diagnostic criteria for a substance- or medication-induced disorder [111, 146]. A positive screen would not close the question, because persistence after a substance has cleared can point to an independent episode [111].

Antidepressant-associated activation is the pediatrician's question. Fluoxetine's labeling directs monitoring for agitation and emerging manic symptoms, and its long elimination half-life and active metabolite mean that exposure falls slowly after the drug is held [341]. The dose was raised about three weeks before the episode began: timing consistent with a drug contribution, and equally consistent with an episode that would have arrived anyway in a sixteen-year-old with two depressive episodes and an aunt with bipolar I disorder. By convention, a full syndrome that persists beyond the physiological effect of the drug is regarded as bipolar disorder rather than a medication effect [111]; with fluoxetine, that judgment takes weeks.

A psychotic disorder was addressed above: nothing observed or reported places an experience outside the mood state.

Depressive antecedents and one figure

In a prospective study of childhood-onset dysthymic disorder, 13% of the children developed bipolar disorder during follow-up [18]. In a birth cohort followed back from adult diagnosis, most adults with a mental disorder at twenty-six had received a juvenile diagnosis, often of a different disorder [59]. Neither makes Simone's two depressive episodes and "gray since thirteen" a prediction; each says that depressive beginnings in childhood are a known antecedent of bipolar disorder. Depression in young people presents with irritability as well as sadness [14], so the sibling fights during her depressive episodes carry no diagnostic weight now. In the NCS-A, the lifetime prevalence of bipolar I or II disorder among adolescents aged thirteen to eighteen was 2.9% [53]. Outpatient diagnoses of bipolar disorder in youth rose sharply in the decade before the Leibenluft and Rich review [66]. The base rate and the rise are both reasons to describe an episode precisely before naming it.

What this case cannot establish

The presentation cannot establish whether the episode is manic or hypomanic in the diagnostic sense: twelve days exceeds the one-week minimum for mania, and the remaining question is whether impairment is marked [111] in an adolescent who attends school and has also spent her savings, disrupted the art room, and stopped sleeping. It cannot establish whether fluoxetine precipitated the episode, whether the weekend at fourteen was a prior hypomanic episode, what her baseline attention is, whether her certainty will stay revisable, or what happens next. It establishes that several functions changed together, over a defined interval, away from her own baseline, and that two pairs of look-alike observations have been separated for the later chapters to use.

Figures for this chapter

Functional neuroscience

Functional layer

The same operations as Jonah's, with different settings

To guide action, the brain must represent a possible outcome before it occurs, relate it to an action, and integrate expected benefit with cost. Anticipatory reward responsiveness (Process) is the motivational response before receipt; hedonic generation (Process) is the pleasurable impact during receipt; effort-cost computation (Process) estimates the demand of an action; effort–benefit integration (Process) combines demand with expected benefit; action initiation (Process) is the transition into performing the selected action.

In the Depressive Disorders module (Jonah), these operations explained why he could care about music and still not open the case. Simone in September shows the same operations with different settings. Nearly every opportunity generates strong anticipatory responsiveness: a page, a mural, prints, a studio, an editor. Effort cost is estimated as small; she has slept three hours and does not register the demand. Integration favors action almost everywhere, and initiation is immediate. Three murals were begun and none finished, because the next opportunity outcompeted the current one before it ended.

Two State parameters name what changed. Reward sensitivity is how much anticipated benefit a given opportunity generates. Approach-motivation gain is how strongly that anticipated benefit is converted into action. Neither is a new capacity; both are settings on operations Simone has always had, treated here as State because they moved together, for twelve days, away from her own baseline. The behavioral approach system hypersensitivity model of bipolar disorder proposes something close to this: heightened sensitivity of approach motivation to reward-relevant events, escalating into hypomanic or manic states under goal-striving conditions [147]. The model rests on group findings under defined conditions; it supplies a vocabulary for Simone's two weeks and does not measure her gain.

In an activated episode ordinary stopping points lose influence. A stopping point is an effort–benefit decision run in reverse: continuing has a cost (dinner is waiting; it is three in the morning) and stopping has a benefit. When gain is high, stopping loses every comparison, and a stop imposed from outside registers as a loss. "You're stealing this from me" is an accurate report of that computation, and it is why the dinner blowup resolved into expansiveness once the pursuit resumed rather than into the shame that followed Jonah's homework reminders.

More activity is not more effective action

More activity does not imply more effective action or greater experienced pleasure. Simone's drawing skill is a Capacity and appears unchanged: forty pages exist and her teacher recognizes the output. What changed is allocation. Attention is captured by each new opportunity, and persistence on one task is what fails. Whether the pages are good is a question the teacher can answer and Simone at present cannot, because her evaluation of her own work runs on the changed settings. This is capacity versus access in its activated form: not less access, but access reorganized so that the capacity is spent on whatever is most recent. Whether she is enjoying it is a further question; increased drive can accompany painful affect, and the irritable intervals show the state is not uniformly pleasant.

Value updating that runs too fast

Pleasure during an event and learning from it are distinct. A simple learning rule shows the difference:

New expected value = old expected value + learning rate × (experienced outcome − expected outcome).

The expression in parentheses is a prediction error. In the Depressive Disorders module (Jonah), a rehearsal expected to be worth two units delivered six, and with a learning rate of one quarter the next expectation became three: one good experience need not overturn a long-held expectation.

Run the rule with Simone's September settings and three things can move at once. The experienced outcome can be inflated: a good page registers as ten. The learning rate can be high: near one, a single outcome replaces the old expectation outright, and "the book" goes from two to ten on one page. And outcome credit assignment (Process), which links an outcome to its antecedent, can widen: a good page is credited to herself as an artist, to the whole book, to prints, to a studio, and to an editor's probable reply. Value updating (Process) then revises all of them at once. The numbers are illustrative; nothing in Simone's presentation measures a learning rate.

Updating in the other direction looks slower. The clinician's doubt was a negative prediction error offered in conversation, and "Maybe. I don't think so" is what it looks like when such an error is registered and barely applied. In a euthymic cohort with bipolar disorder, a reduced positive-versus-negative updating bias predicted earlier relapse [148]. The finding does not show faster positive updating in every mood state, and Simone's asymmetry is inferred from one exchange. A signal that changes with surprise teaches an organism about opportunities; it is not a meter of how happy the organism feels [326]. A fast updater is not happy because it is fast. It is certain because it is fast.

Representations of self and of the future

A self-model (Representation) is what is held about who one is; a future model (Representation) is what is held about what is coming. A month ago Simone's self-model contained "gray." Now it holds "I finally have the energy I was supposed to have," and the future model holds a published book, a studio, and an editor's reply. Fast, wide updating writes into these representations, and they feed back into anticipatory responsiveness: each new opportunity is evaluated against a self-model in which she succeeds.

What separates this from the beliefs in the Psychosis module (Ravi) is the process, not the content. The representation is still revisable; she can hear doubt and locates her certainty inside herself. A self-model that stops taking evidence, or a future model populated with events others cannot see, would be a change in process and would be recorded as one. Recovery will also rewrite the self-model, not necessarily back to gray. What Simone thinks of the emails and the $340 when the gain returns to baseline is itself a clinical event, to be anticipated rather than discovered.

Sleep, circadian state, and their coupling to activation

Homeostatic sleep pressure accumulates during wakefulness and dissipates during sleep; circadian timing regulates the daily pattern of sleep propensity in relation to biological time. Observed sleep timing and endogenous circadian phase provide different information, and a bedtime alone does not measure phase [328].

In May the account works as written: pressure accumulated, was felt as exhaustion, was resisted with caffeine, and was repaid on Saturday. In September the felt need has come apart from the hours: twenty hours awake, and "not tired." Sleep need (State), the felt requirement for recovery at a given accumulated pressure, is the variable that moved, and it moved with the others. Whether pressure accumulates and is not felt, whether an arousal signal overrides it, or whether her circadian phase has shifted cannot be established from a report.

The coupling runs both ways. Sleep loss precedes and can trigger manic episodes in some patients, and reduced sleep during an episode is proposed to feed further activation [149]. Disruption of the social rhythms that entrain the circadian system has been proposed as one route to episode onset, and stabilizing those rhythms has been tested as treatment [150, 151]. Sleep is therefore both a target and a measurement. The four fields that separated May from September (when sleepiness arrives, bedtime, waking, whether recovery is needed) are a monitoring instrument [149]. The return of felt sleep need is a clinical observation to record, not a validated early sign that the state is changing.

Why an episode is a State and not a trait

A State is a condition that changes access to capacities. Simone's capacities (drawing, sequencing her project, hearing doubt, weighing evidence) are present; what changed are the settings under which they are accessed, and reward gain, sleep need, thought speed, and the threshold for stopping all moved together, for twelve days, away from her baseline. That coordinated, time-bounded shift is what an episode is. A trait is present across states and time. The fights with Ollie are close to a trait of that relationship, present in depression, now, and in the weeks between. The dinner irritability is State-bound: it appears when high gain meets an external stop and dissolves in minutes.

This is the distinction the pediatric bipolar literature has argued over. Leibenluft and Rich set the narrow phenotype, distinct episodes of elevated mood or grandiosity against a baseline, apart from chronic irritability with hyperarousal and no episodes, and summarized longitudinal and family data suggesting the two follow different courses, with the chronically irritable group at risk for later depressive and anxiety disorders rather than mania [66]. DMDD, which followed from that work, is defined by chronic irritability and excludes a distinct period of elevated mood lasting more than a day [7]. Simone's case is on the episodic side. That describes the shape of her presentation; it does not predict her course.

A single episode establishes a syndrome and its timing. It cannot establish whether fluoxetine precipitated it or whether it would have occurred anyway; the polarity or frequency of any future episode; whether the depressive episodes at thirteen and fifteen were first expressions of the same disorder, which the 13% figure from childhood-onset dysthymia [18] and the heterogeneous juvenile antecedents in a prospective birth cohort [59] make possible without making likely; which parameters return to baseline first; or any mechanism.

Circuits, development, and genes

Reward circuits and what a signal means

The ventral striatum participates in processing reward-predicting information and its motivational significance; cortical regions contribute representations of outcomes, context, goals, and costs; dopaminergic input influences activity and learning within these circuits. In 1,576 adolescents, reduced ventral striatal responses during reward anticipation were associated with anhedonia [323]. In the activated direction, reward-circuit findings in bipolar disorder are heterogeneous; in youth at familial risk they include frontal activation and connectivity differences, not a consistent increase in striatal or orbitofrontal anticipation responses [152, 153]. In healthy adults, dopamine biased the weighting of benefits against cognitive effort costs [324]: a plausible route to altered effort–benefit integration that does not show Simone's dopamine is elevated, or explain why a graphic novel rather than a sport or a fight became the pursuit.

Timing, cells, and the state of the receiver

In mice, disruption of CLOCK produced changes described as mania-like, including altered activity and reward-related behavior: an experimental link between molecular timing and behavior in an animal model, not evidence that pediatric bipolar disorder is a CLOCK mutation syndrome [329]. It is the level at which sleep, timing, and reward behavior are seen to share machinery, which makes the coupling above more than an analogy.

The N-methyl-D-aspartate receptor's response depends on membrane voltage, so the effect of a transmitter depends partly on the state of the receiving cell [342]. That is the cellular version of the principle the functional layer uses: a State changes what an input does without changing what the system can do. It is also why a molecular event can change the effectiveness of communication within an already learned network, yet not explain the individual outcome.

Development and genetic liability

In a meta-analysis of population studies dating onset by first symptoms, diagnosis, or hospitalization, the median age at onset of bipolar and related disorders was 33 years [154]. Simone's episode began at sixteen, well before that median, in a period in which reward-related striatal systems are described as relatively mature while prefrontal regulatory systems are still developing [155]. Each is a plausible reason for timing; neither has been shown to be the reason for any individual.

The 2025 bipolar genome-wide analysis of 158,036 cases and roughly 2.8 million controls established a broad architecture of liability whose findings varied with ascertainment and subtype; it does not establish the cause or diagnostic significance of an individual adolescent's activation [332]. Risk is elevated in first-degree relatives of affected people [156]. Simone's aunt is a second-degree relative and her grandfather was never assessed; the history raises the prior without specifying it. Lithium response has been reported to cluster in families [157]; her aunt's years on lithium are a family fact, not a pharmacogenetic test.

Moving between levels of evidence

A reward task tests a process under defined conditions; a learning rule shows how parameters could combine and measures none of them; a mouse with a disrupted clock gene demonstrates shared machinery; a genome-wide study describes a distribution of liability. Simone's forty pages, three-hour nights, and thrown brush sit at none of these levels.

What this case cannot establish

Nothing in Simone's presentation measures reward sensitivity, approach-motivation gain, a learning rate, sleep pressure, or circadian phase; those are functional descriptions and group-level findings that give those two weeks a vocabulary and her clinicians questions. Her family history and her own depressive episodes are recorded antecedents, not established contributors. Her thrown brush cannot establish a neural event. What the functional account establishes is a set of distinctions treatment can act on: gain versus capacity, felt sleep need versus hours slept, a revisable self-model versus a fixed one, and a State that moved several settings together versus a trait that was there all along.

Figures for this chapter

Treatment

What Simone, her mother, and the pediatrician are asking for

Simone wants to keep the energy and finish the book. Her mother wants her to sleep, wants to know whether she is safe, and wants to know whether this is the medicine. The pediatrician wants to know what to do with the fluoxetine. These goals differ, and Simone's is partly a product of the state being treated. The plan serves all three while the diagnosis stays open: the first decisions concern safety, setting, and the drug already in her body, and the diagnostic question is answered by observation over the following weeks.

First decisions: safety, setting, and the fluoxetine

When sustained activation emerges during antidepressant treatment, the clinician characterizes sleep need, thought speed, behavior, and the temporal relation to treatment promptly; agitation alone should not be called bipolar disorder, and a clear manic syndrome should not be treated as ordinary recovery. Management depends on severity and the complete presentation, including safety and the role of current medication.

For Simone the safety questions are specific: money, nighttime messaging to strangers, driving, sexual risk, and any plan that takes her out of the house at night. Nothing observed is dangerous in itself, she is not psychotic, and her family can supervise, so outpatient care with review within days is defensible. A belief that stops yielding, an experience located outside herself, an unsafe act, or a family that can no longer keep her in would change that.

The pediatrician has held the fluoxetine, which is the right first step and does not by itself answer the question. The label directs monitoring for agitation and emerging manic symptoms, and its long elimination half-life and active metabolite mean that exposure declines slowly after the last dose [341]. "Held" is not "gone." The window in which the episode could fade with the drug is weeks long, and a syndrome that persists at full intensity beyond it is treated, by convention, as bipolar disorder rather than a medication effect [111]. The drug stays held, is not restarted for the depression that may follow, and the timeline is written down now while it is recoverable.

Antimanic treatment: episode-specific evidence

Antimanic efficacy and treatment of bipolar depression are different evidence questions, and neither is answered by the adolescent depression trials in the Depressive Disorders module (Jonah). For acute mania in adolescents, placebo-controlled trials support several second-generation antipsychotics, and lithium has pediatric trial evidence [158, 159]; these are acute-phase findings, separate from maintenance and long-term safety. A head-to-head trial in children aged six to fifteen, younger than Simone, found higher response with risperidone than with lithium or divalproex, with greater weight gain on risperidone [160]. Selection among antipsychotics should weigh efficacy alongside differing adverse effects; weight gain and metabolic change are prominent in adolescents, and monitoring of weight, glucose, lipids, prolactin, sedation, and movement is part of prescribing [158, 145].

Lithium's therapeutic mechanism is incompletely understood, while its renal elimination and narrow therapeutic range have immediate prescribing implications: hydration, intercurrent illness, interacting medication, thyroid and renal function, and measured exposure all require attention [341]. Current US labeling for lithium covers acute manic or mixed episodes and maintenance in bipolar I disorder from age 7, so it covers Simone; the label is still read before prescribing [161]. Her aunt's years of wellness on lithium will weigh with the family; familial clustering of lithium response has been reported [157], and it raises a prior without being a test. The choice between lithium and an antipsychotic rests on adverse-effect profiles, her physiology, the family's capacity to manage blood levels, and her own view once she can hold one.

If Simone is prescribed an antipsychotic, the functional account offers a vocabulary and no more. Dopamine D2 antagonism could be described as turning down the reward gain the previous chapter named, and the dopamine findings make that a plausible route. The trials did not measure gain; they measured mania rating scales, and they do not show that reduced gain mediated the ratings or which adolescents will respond. A receptor action can change the effectiveness of communication within an already learned network, yet this does not explain the individual outcome, undo the emails, or decide whether the episode was the drug or the disorder.

Bipolar depression, if it follows, is a separate evidence question. In a pediatric trial, lurasidone improved depressive symptoms over six weeks in bipolar I depression; a receptor list cannot establish which action produced benefit, food substantially affects exposure, and efficacy for mania is not established in the label [336, 341]. The evidence for adolescent unipolar depression does not transfer: fluoxetine is the drug that was held.

Sleep as treatment and as measurement

Sleep need is the State variable the family can observe most directly, and the plan treats it and measures it at once. Energy drinks stop; nicotine is discussed; a fixed wake time, morning light, and a dark room from a fixed hour are set with Simone rather than imposed, because an imposed stop registers as a loss and the family has seen what that produces. Short-term pharmacological help with sleep is considered on its own evidence, drawn largely from adult guidance, if the schedule alone fails [162]. Interpersonal and social rhythm therapy, which stabilizes the timed cues of daily life, has randomized trial evidence in adults with bipolar disorder; its adolescent adaptation has only preliminary results from a small open trial [151, 163].

Her mother's log is upgraded to the four fields that separated May from September: when sleepiness arrives, bedtime, waking, and whether recovery is needed. The fields monitor changes in sleep duration and subjective sleep need [149]; a return of felt need is a clinical observation, not a validated early marker that the episode is ending. Kept after recovery, the same fields become the early-warning instrument for the next episode, and will distinguish a future finals week from a future September.

Psychosocial treatment acts on the processes named in the formulation

Family-focused treatment for adolescents with bipolar disorder, combining psychoeducation, communication training, and problem solving, has been tested in a randomized trial. It produced faster recovery from depressive symptoms and less depressive morbidity, but did not significantly shorten recovery from the index mood episode or delay recurrence overall [164]. Its content maps onto this family's questions.

Psychoeducation gives the family the state-versus-trait distinction of the previous chapter. The fights with Ollie are not the illness and need not become its barometer. The dinner blowups are the state meeting an external stop, a different problem with a different solution: stopping points agreed in advance, at hours Simone has accepted, rather than interruptions at the peak of a page. Ollie is told plainly what is happening.

The graphic novel is not confiscated. The forty pages are real, her drawing is a preserved Capacity, and removing the project would remove the activity that could later carry her out of a depression, as the duet did for Jonah. What is contained is the judgment around it: money is held by her mother, messages to strangers wait until morning, and the editor's email, which cannot be recalled, is planned for, because a reply or a silence will arrive as a negative prediction error and the state it lands in matters. The art teacher keeps three fields: interrupts, starts new work, finishes.

The therapist from last year is re-engaged. Cognitive work waits until Simone can hold a doubt longer than an exchange; when she can, the material is the self-model of the previous chapter, and the aim is a representation of these two weeks complete enough that what emerges is neither "I was insane" nor "I was finally myself."

The depressive side, and what she will think of it

The depressive side of this disorder is where the danger is. In longitudinal follow-up, depressive and mixed symptoms make up a substantial part of the course in young people with bipolar disorder, and suicidal thinking and behavior are common [165, 166]. Simone said that for the first time in years she does not want to die. Her mother's fear is well founded for a reason she may not have articulated: the wish is likely to return when the gain does, and to a self-model that now contains two weeks she may be ashamed of. The safety plan is written now, before she needs it, and names the sleep fields, the mood change her family will recognize, and who she tells.

After the episode: what a single episode leaves open

Continuation treatment after a first manic episode is standard practice, but the evidence for how long adolescents should continue, and on what, is thin [167, 168]. The decision is made with Simone at a review in which she can weigh it, the family history and two prior depressive episodes on one side and the adverse-effect burden of long-term treatment on the other. The review questions are the formulation's: has felt sleep need returned; has the gain come down before the drawing has stopped; does she still hear doubt; what has the self-model become; has the depressive side arrived. If the episode faded with the fluoxetine and she stayed well, the formulation changes and what she was treated for is reopened honestly.

What this case cannot establish

The plan cannot establish that fluoxetine caused or did not cause the episode, that any antimanic agent will suit Simone, that a receptor mechanism explains her response, or that sleep regularization or family treatment acts through the functional process it is described as targeting. It cannot establish whether she will have another episode, or of which polarity. What it establishes is a set of treatments justified by trial evidence for the episode she is in, protections for her judgment that leave her capacity intact, a monitoring instrument built from her family's own observations, and a defined list of signals that would change the plan.

Figures for this chapter

---