‘My mood is better so why am I struggling so much?’: Cognitive Challenges in Bipolar Disorder

Patients with bipolar disorder often manifest cognitive disturbances during acute manic, depressive, and mixed states which include difficulties with attention, concentration, planning and memory.   It may come as a surprise, though, to learn that cognitive deficits can persist into euthymia – a state of relative mood stability.   Indeed, approximately 40 % of euthymic bipolar patients show evidence of cognitive impairment (1,2) involving attention, executive function (i.e. planning and organization, cognitive flexibility and set-shifting, and working memory), verbal memory, processing speed and visual memory (3,4).  Patients with bipolar I and bipolar II disorder appear to have a similar pattern of deficits (4).   The purpose of this brief review is to identify cognitive deficits that help distinguish bipolar disorder from other disorders of mood, memory, and executive functioning.   An improved understanding of these deficits will help clinicians tailor their treatment interventions to address the individual needs of patients.

How do such deficits affect the day-to-day functioning of a patient with euthymic bipolar disorder?  One example would be a college student who is struggling to perform well in school despite having fully recovered from a manic episode several months before.   Areas of difficulty include maintaining focus on class lectures, learning new material, and completing coursework on time.   Another scenario involves a manager, back at work following a prolonged depressed period, who feels completely overwhelmed in the course of a seemingly normal day.  Routine tasks take much longer and lead to an overall feeling of being ‘stuck,’ unable to move on to other projects or work on multiple projects simultaneously.

This constellation of symptoms described above resembles another disorder of executive functioning, Attention Deficit Hyperactivity Disorder (ADHD).   Research on cognitive findings in bipolar disorder has had to address the question of which psychological test (s) can most effectively differentiate between bipolar disorder and ADHD.  In a 2014 study, conducted by Silva et. al. (5), patients were divided into three groups:  ADHD with bipolar disorder (n=51), ADHD without bipolar disorder (n=278) and healthy subjects (n=91).  The Wisconsin Card Sort (WCST) was administered to all of the patients.   In the WCST, patients are asked to organize cards displaying various colors and shapes in the face of changing rules.  The test is a measure of the patient’s ability to ‘set-shift,’ or adapt to shifting contingencies.  Patients with both ADHD and bipolar disorder had lower scores than patients with ADHD without bipolar disorder and healthy controls (the latter two groups did not differ from each other).  The authors concluded that impairments in ‘set-shifting’ are strongly related to bipolar disorder but not ADHD (5).

Several years later, Gruber et. al. administered the Multi-Source Interference Task (MSIT) to 29 patients with euthymic bipolar disorder and 21 healthy controls with concurrent functional magnetic resonance imaging (fMRI) (6).  The MSIT consists of three-digit stimuli sets (using numbers 0, 1, 2, or 3) that are presented briefly on a screen.  Each set contains two identical distractor numbers and a target number that differs from the two distractors.  Participants are asked to press a button corresponding to the identity of the target number that differs from the two distractor numbers.  There are two scenarios presented:  a ‘control’ condition in which the target number corresponds to position on keypad, and an ‘interference’ condition in which the target number does not correspond to position on the keypad.  Overall, the patients with bipolar disorder exhibited slower response time and lower accuracy in the interference condition, which required adaptation to a new and more challenging situation.  In addition, neuroimaging revealed decreased activation of the anterior and middle cingulate cortex and increased activation of the dorsolateral prefrontal cortex, which are regions of the brain associated with cognitive control processing (6).

Taken together, these studies help us piece together a pattern of neurocognition that is potentially specific to bipolar disorder:  deficits in planning and set-shifting as a result of impairment in cognitive control.   Cognitive control is a form of executive functioning that allows patients to adapt their thoughts and actions appropriately in the face of changing environmental scenarios (6).  More specifically, it is ‘an executive control system… whose central purpose is to overcome (i.e. to resolve) interference or conflict in cognitive control processing’ so that one can ‘maintain adequate performance in the face of significant distraction.’   Interference comes in the form of ‘task-irrelevant information’ that results in slower processing speeds and increased errors (7).   The cognitive control system is what allows us to ‘switch gears’ and move on in a more efficient way when environmental changes occur.

Clinically, these findings point to a difference in how patients with bipolar disorder process new information, including environmental changes, differently from unaffected controls due to alterations in brain connectivity.   Research on neurocognitive aspects of bipolar disorder, while still in its early stages, can assist clinicians in targeting their therapeutic interventions.   In recent years, several cognitive rehabilitation-based therapies have emerged that show promise.

In cognitive remediation (CR), attention, memory and executive functioning are targeted primarily with computerized exercises.   CR is said to work by enhancing the neuronal plasticity of the brain by ‘restitution’ (stimulation of cognition by repetitive exercises) and ‘compensation’ (such as memorization skills and use of environmental aids) (8).   In study of 39 bipolar patients showing cognitive impairment, improvement after CR was observed in working memory (p=0.043), problem solving (p=0.031) and divided attention (p=0.065) (8).    In another recent study, 75 patients with bipolar disorder were randomized to a 70-hour computerized cognitive remediation(CR) program or a computer control.  Post-treatment results showed medium to large positive effects of CR on processing speed and visual memory (9).

Critics of CR argue that while patients may do well on computer-based tests following the intervention, there is no significant improvement in overall functioning or quality of life.   Functional remediation has thus been developed, in order to allow the use of ‘ecological neurocognitive techniques’ that improve outcome in daily life (10).  When tested in a multicenter controlled study involving 77 patients with bipolar disorder, there was an improvement in functional outcome after 21 weeks of treatment, especially in the areas of enhanced occupational and interpersonal functioning.  However, performance on cognitive testing did not significantly improve.  The authors suggest that ‘even though cognitive deficits may persist, patients exhibit greater ability and more strategies to cope with those deficits in daily life’ (11).

More recently, cognitive behavioral rehabilitation (CBR) has been developed and is still under study.  This new approach combines cognitive remediation (CR) with cognitive behavioral therapy (CBT).  This method includes using CBT to identify ‘automatic thoughts’ and ‘thought distortions’ and to restructure these thoughts.  ‘Mental flexibility’ is also addressed (12).  While results are pending, this study is of interest in that it suggests that it may take on some of the core neurocognitive challenges previously mentioned in this article, specifically, problems with ‘set-shifting.’

To summarize, patients with bipolar disorder often experience difficulties with attention, memory, and executive function that persist even in the absence of a mood episode.   These cognitive deficits can cause disruptions in many areas of daily functioning including home, work and school.   A clear pattern of deficits has not yet been identified, but evidence thus far converges on attention, memory, and executive function as being affected.  One finding of interest across several studies is that patients with bipolar disorder struggle with ‘set-shifting’ due to deficits in cognitive control.   These deficits can be localized to the anterior/middle cingulate cortex and its connections to the prefrontal cortex.     Rehabilitation strategies currently being studied consist of computer-based cognitive remediation, functional remediation, and/or cognitive-behavioral rehabilitation.   As we increase our understanding of the cognitive manifestations of bipolar disorder, we will have more choices available to help patients achieve optimal functioning in many areas of their lives.

Susan Stern, M.D.



  1. Martino DJ et al. Heterogeneity in cognitive functioning among patients with bipolar disorder.   J Affective Disord 2008; 109: 149-156.
  1. Iverson GL et al. Identifying a cognitive impairment subgroup in adults with mood disorders.   J Affective Disord 2011; 132: 360-367.
  1. Goldberg JF and Chengappa KNR. Identifying and treating cognitive impairment in bipolar disorder.  Bipolar Disord 2009; 11 (Suppl. 2): 123-137.
  2. Palsson et al.  Neurocognitive function in bipolar disorder: a comparison between bipolar I and II disorder and matched controls.  BMC Psychiatry 2013; 13: e1-9.
  3. Silva KL et al. Could comorbid bipolar disorder account for a significant share of executive function deficits in adults with attention-deficit hyperactivity disorder?  Bipolar Disord 2014; 16: 270-276.
  4. Gruber SA et al. Decreased cingulated cortex activation during cognitive control processing in bipolar disorder.  J of Affective Disord 2017; 213: 86-95.
  5. Melcher T et al. Functional brain abnormalities in psychiatric disorders: neural mechanisms to detect and resolve cognitive conflict and interference.  Brain Research Rev 2008: 96-124.
  6. Veeh J et al. Cognitive remediation for bipolar patients with objective cognitive impairment: a naturalistic study.  Int J Bipolar Disord 2017; 5: e1-e13.
  7. Lewandowski KE et al. Treatment to enhance cognition in bipolar disorder (TREC-BD).  J Clin Psych 2017: 78; e1-e8.
  8. Martinez-Aran A et al. Functional remediation in bipolar disorder.  Clin Pract & Epidem Mental Health 2011; 7: 112-116.
  9. Torrent C et al. Efficacy of functional remediation in bipolar disorder: a multicenter randomized control study.  Am J Psychiatry 2013; 170: 852-859.
  10. Gomes BC et al. Cognitive-behavioral rehabilitation vs. treatment as usual for bipolar patients: study protocol for a randomized controlled trial.  Trials 2017; 18: 142.

Mood and Affect: Definitions and Basic Concepts

If you’ve come to this website and are reading this article, the odds are good that you or someone that you know has been diagnosed with a mood disorder.  You might be seeking more information about emotional illnesses.   But before diving into any of the numerous areas of research presented at our site, what if we pause and ask a very basic but important question:  what are moods?  We use this term all the time, but do we truly understand what they consist of and why we experience them?  Even more fuzzy, what about the term, ‘affects’?  We specialize in the treatment of ‘affective illness’ here, but what is an affect and how is it different from a mood?  This brief review will summarize a generally accepted concept of emotions, describe their composition and functional purpose and present current research in this area.  To start, let’s begin by taking a look at how these terms are currently used in clinical practice.

Current thinking defines mood as an emotional state with relative persistence over time.  Affects, in contrast, are shorter, more reactive emotional experiences.  Psychiatry and psychology trainees all learn how to describe affect in terms of degree (standard vs. reduced, blunted or flat), range (full vs. restricted) and congruence (does it fit with what the person is talking about at that moment?).  The problem here is that the definitions are circular – mood and affect are explained by emotion, and emotion is referenced to mood and affect – and the descriptions are limited.  It doesn’t really help us understand the true nature of mood, affect or emotion.  Greater insight into this problem is achieved only by turning to other areas of science that put human emotional experience into perspective:  evolution, child development, animal research, and temperament.

In his book, The Expression of the Emotions in Man and Animals, Charles Darwin made several observations that laid a foundation for the scientific study of emotion [1].  First, he noted that facial and behavioral expression of emotions were similar between species, especially between non-human primates and humans.  In response to a predator, for example, dogs, chimps and humans would all react with raised eyebrows, widely open eyes, hair standing on end (piloerection) increased heart rate, heightened muscular tone, and behavioral withdrawal.  He thus reasoned that the emotional response system in mammals is genetically inherited and evolutionarily ancient.  Second, he proposed that emotional expression consisted of particular states of mind that were combined with defined patterns of activity.  Fear was coupled with withdrawal and the above autonomic nervous system changes (increased heart rate, heightened blood pressure, cortisol secretion, etc..).  Love was linked to behavioral approach.  This identification of a motor (i.e., movement) component of emotional experience was huge.  Last, he suggested that human facial displays of emotion are the vestiges of evolutionarily earlier, stimuli-specific, whole-body response patterns.  So instead of fleeing in response to all threats, humans have learned to initially respond with a facial display of fear; instead of attacking when we are angry, we manifest only pursed lips and a tightening of certain other facial muscles.  Inherent in these descriptions is the idea that emotional displays, like the older, more global behavioral response patterns from which they derive, are also adaptations to the environment.  Survival also moves us to seek out nourishment and love from our environment.   Darwin’s contribution to the study of emotion cannot be overstated.  He recognized its inherited nature, its subjective/psychological, autonomic and motoric components, the reactive quality of these psycho-motor patterns to specific, environmental stimuli, and the communicative and adaptive purpose of emotional experience.

Approaching this subject from a developmental and cross-cultural perspective, psychologist Sylvan Tomkins has studied facial expressions in early childhood [2].   After reviewing thousands of facial expressions in newborns from many different cultures, Tomkins concluded that the face is, in effect, the primary organ of emotion.  It is the first responder and, as such, guides and leads the autonomic, other motoric, and subjective elements identified by Darwin.  Furthermore, just as Darwin noted similar behavioral emotional responses among all mammals, Tomkins found that there is a discrete number of common facial configurations used across all human societies.  Whether an African tribesman or an urban Asian baby, each person shares the same basic vocabulary of facial expressions.  He identified nine basic facial expressions that are innate, that is, present at birth, and found throughout the world.  To really bring this home, check out the Tomkin’s Institute website that shows this universal language of the face: (http://www.tomkins.org/what-tomkins-said/introduction/nine-affects-present-at-birth-combine-to-form-emotion-mood-and-personality).  These include two positive forms (interest-excitement and enjoyment-joy) and seven negative varieties (distress-anguish, anger-rage, fear-terror, shame-humiliation, disgust, and others).  Complementing Darwin’s work, Tomkins believed that these innate facial displays are automatic responses to environmental stimuli.  Further, he suggested that these early, automatic reactions form the foundation for later emotional experience.  In other words, the involuntary look of terror that comes on an infant’s face in response to a barking dog is the first component of what will become that young child’s and later adult’s experience of anxiety.

The extraordinary research of veterinarian scientist, Jaak Panksepp, grounds the theories of Darwin and Tomkins in neurobiology [3, 4].  In work that spans a forty-year career, Panksepp used electrical stimulation and lesion methods (kill certain brain cells and study behavioral effects) to identify the discrete affects experienced by a variety of mammals and their distinct neurobiological underpinnings.  Electrically stimulating one pathway produces fear and withdrawal; buzzing another results in rage; chemically lesioning a third results in the elimination of a specific emotional reaction.  His findings converge with and enhance the evolutionary and developmental perspectives.  Like Darwin and Tomkins, Panksepp’s animal work confirms the existence of a finite number of affective behavioral reactions.  Using capital letters to identify them as systems, these included: LUST, CARE, PANIC, PLAY, FEAR, RAGE, and SEEKING.  These systems have a rough, but imperfect correspondence to the core facial expressions put forward by Tomkins.  More interesting, the circuit mapping revealed that each of these primary affects was associated with their own individual neural pathways, often using specific neurotransmitters within that pathway.  Panksepp’s animal research establishes the hard-wired, neurobiological foundation of affective experience.  Thus, affects take shape as inborn, composite reaction patterns, set off by, and specific to, particular environmental stimuli, which are mediated by evolutionarily old, emotion-specific brain pathways.  We now have a much richer explanation and description of affects.  What about mood?

Freud, and more recently Michael Basch, wrestled with the issue of moods and their relationship to affects [5, 6].  Both considered moods to represent a more refined, developmentally mature expression of affects.  Basch, in particular, posited a developmental line of affective experience that begins with the type of brief, affective reactions described earlier, progresses to emotional experiences that reference an emerging sense of self, and culminate in the capacity for empathy.  Thus, developmentally, emotions proceed from a primal, unreflective, “ouch!” and withdrawal in response to pain, to a more personalized, “that hurts me” level of expression, and finally to the empathic realization that others experience pain in similar ways to oneself.  In parallel with this developmental progression, the time course of emotions begins to extend and longer-lasting mood states develop.  These longer mood stretches begin to color and influence our personalities.  When repeated reactions of fear linger, we become anxious individuals.  Affective reactions of sadness that recur and extend into frequent mood states begin to impart a blue tone to one’s personality.  That sequence makes sense.  But what of the reverse direction?  Can attributes of personality shape the nature of our emotional inclinations?   Experimental studies in this area form the last section of this report.

The research of Richard Davidson, a psychologist at the University of Wisconsin – Madison, was prompted by individual differences in emotional expression.  Why do some people incline towards brighter, more outgoing moods and others, towards a more brooding and pessimistic style?  Davidson suggested these differences arise from one’s underlying, affective style – this is his term for a broad dispositional temperament that biases individuals towards specific affective and mood experiences [7].  Affective style describes an individual’s tendencies to respond to particular stimuli in specific ways.  For example, mild surprises might typically elicit laughter in one person, but a startle reaction in another.  These characteristic reactions can be measured based on magnitude (how much does one laugh or get startled?), duration, threshold (how large a stimulus is necessary to prompt the reaction?), peak to rise (does the reaction jump up quickly or does it percolate upwards?), and recovery time.  This breakdown of affective responses into its constituent components was termed affective chronometry – a truly fascinating field in its own right.  Now, let’s go back to the topic of affective style.

It turns out that affective style correlates broadly with two fundamental neurological orientation systems that cause us to respond in habitual ways towards environmental events.  The operation of the Behavioral Activating System (BAS) and the Behavioral Inhibition System (BIS) – recently discussed in depth by Dr. Stern in her blog at this site – act as motor biases that slant us towards and away from rewarding and aversive stimuli respectively.  Think about it like posture:  imagine one person with an upright, erect stance who handles things in an assertive, proactive way and another with hunched over shoulders and a slouch who is waiting for life’s next disappointment.

The BAS moves us towards desired goals and favored outcomes, the BIS causes us to recoil from threats and losses.   High BAS individuals are prone to extraversion and impulsivity.  Impaired BAS system functioning is associated with the resignation and the joylessness of depression.  High BIS activity is associated with anxious apprehension and avoidance, whereas deficient BIS would present as disinhibition and recklessness.

Davidson first demonstrated that affective style correlates with the activity of these two fundamental motivational systems.  His later research had two blockbuster findings.  First, the optimistic, challenge-embracing, high BAS orientation is associated with increased left frontal brain wave activity that is measured by EEG (electroencephalogram; a device which measures brain waves).  Conversely, the fearful, depressed high BIS inclination matches up with high right frontal EEG activity.  Therefore, whichever side of the front part of our brain is more active appears to shape our outlook and dispositional mood.  Second, these brainwave activity asymmetries are present from birth onward.  Davidson found, for example, that 10 month old infants who cried in response to maternal separation had higher right than left prefrontal EEG activity.  Not only are these differences present from an early age, they are stable over time.

Thus, we are born with a basic, hard-wired, orientational preference to respond to some events more so than others, and in particular ways.  These neurobiological biases predispose us towards certain affective reactions.  Affect is the rapid, automatic, involuntary and adaptive response to some environmental occurrence.  The repeated experience and lingering persistence of these higher-probability reactions generates the more developmentally mature, longer-term moods that we experience.  Finally, the persistence of the mood states themselves invariably gives color to our emotional personalities.

Moods are complicated.  This review has attempted to put them into perspective through the use of evolutionary, developmental and neuroscience research.

John Gottlieb

  1. Darwin C: The Expression of the Emotions in Man and Animals. NY: Appleton; 1872.
  2. Tomkins S: Affect Imagery Consciousness, vol. 1 and 2. NY: Springer; 1962-3.
  3. Panksepp J, Panksepp J: Affective consciousness: Core emotional feelings in animals and humans. Conscious Cogn 2005, 14(1):30-80.
  4. Panksepp J, Panksepp J: Affective neuroscience of the emotional BrainMind: evolutionary perspectives and implications for understanding depression. Dialogues in Clinical Neuroscience 2010, 12(4):533-545.
  5. Solms M, Nersessian E: Freud’s Theory of Affect: Questions for Neuroscience. Neuropsychoanalysis 1999, 1(1):5-14.
  6. Basch MF: The Concept of Affect: A Re-examination. Journal of the American Psychoanalytic Association 1976, 24(4):759-777.
  7. Davidson RJ: Affective Style and Affective Disorders: Perspectives from Affective Neuroscience. In., vol. 12: Psychology Press; 1998: 307 – 330.


Understanding The Therapeutic Action of Lithium: The Brain as Complex Real Estate



The Lithium Membrain by Anne Naylor.  Lithium coursing through veins in the brain (blue) provides mood stability by acting as a membrane that prevents the effects of the various faces of the illness (circles) on the brain (neural networks).

(Copied with permission from: Anne Naylor and Malhi, G. S. “Lithium therapy in bipolar disorder: a balancing act?” The Lancet 2015 386(9992): 415-416.)





Metaphors are helpful short-hands that enable us to visualize complicated relationships in simple, schematic ways.  The monoamine theory of depression suggested a basic deficit model of catecholamine neurotransmitter dysfunction in depression:  simply too little norepinephrine and /or dopamine [1].  Antidepressants worked by increasing the levels of these communication molecules.  Using this theory, you can almost see the low markings on the oil dipstick or the gas tank gauge and know that a fill-up will solve the problem.   This visual depiction of depression has had a long, useful run and has helped many patients envision their illness and supported their willingness to use medications to fight these chemical shortfalls.

But what of lithium and its role in manic depressive illness?  What’s a simple picture of bipolar pathophysiology and the beneficial effects of lithium?   What do we tell our patients when they ask:  how does lithium work?

In confronting this question, it is immediately apparent that the situation is more complicated with bipolar disorder than with depression.   First, bipolar disorder is simply more pleomorphic than unipolar depression.  It includes a range of pathology from manic to depressive and mixed states, the increased propensity for relapse, the intrinsic switch process between states, the higher levels of persistent subsyndromal symptoms, and the increasingly recognized cognitive impairment and neurodegenerative changes associated with disease progression.

In keeping with this more complicated picture, lithium has multiple effects on manic depression including acute anti-manic properties, slower antidepressant activity, prevention of both manic and depressive relapse, anti-suicidal effects, and neurotrophic and neuroprotective impact.  Each of these actions proceeds with a distinct time course.  This multiplicity of effects tells us that any simple, up-down or deficit-correction model of lithium’s mechanism of action in this disease is bound to be grossly misleading.

This report draws on three recent articles that review the pathophysiology of bipolar disorder and related data about how lithium works its magic [2-4].  While approaching this topic from slightly different angles, the shared themes and overlap of these works far outweighs their differences.  In addition to these three review articles, I’ll also reference an utterly fascinating new experimental study from Nature that highlights lithium’s novel neuromodulatory role [5].   The overall objective here is to achieve a simple, coherent narrative that explains this unique ion’s mechanism of action.

Given that lithium acts at multiple different levels throughout the nervous system, we’ll start at the cell membrane, the basic border which supports the overall structural and functional integrity of the cell.  From there, we’ll move to cellular communication, starting with intercellular, neurotransmitter-based signaling.  We will then proceed to the intracellular translation or transduction of those signals via internal, second messenger systems; this will lead us to the ultimate target of cell signaling which is the alteration of the cell’s genetic instructions, it’s DNA, through the action of transcription factors.  Finally, we’ll move from the individual cell to take a look at larger, more macroscopic aspects of brain structure in bipolar disorder.  Each step of the way, we’ll examine the role of lithium in addressing these different levels of dysfunction.  To help us understand these complex processes, we will refer to the metaphor of the brain as a complex estate, requiring the work of many to maintain its upkeep.


  1. Intracellular electrolyte balance. This involves maintaining the right gradient of ions –sodium, potassium, calcium and others – between the intra- and extracellular spaces; the inside and outside of the cell.   Proper gradients are necessary for optimal functioning of the cell.  Imagine this like the walls of a house and the maintenance of air flow.  The ideal structure will contain the optimal number and size of windows and doors to allow ventilation that is not too easy but not too difficult.   Too many openings and the house won’t provide adequate containment; too few, it will become a hermetic cave.    Some of the earliest and most replicated findings on pathophysiology in bipolar disorder describe disturbed ion gradients – specifically, increased intracellular sodium and calcium –  in the neurons of bipolar patients resulting in impaired activity.   These ion gradations are maintained through the action of special chemical pumps located in the cell membrane which act to shift ions into or out of the cell.  Now it gets complicated:  the operation of that cell membrane pump requires energy.  Energy to power those pumps comes from an intracellular organelle called a mitochondrion, which is a small energy factory located in all cells.  When the mitochondria function normally, the cell has the energy to shift these ions at the appropriate rates to achieve and maintain optimal ion gradients.  When mitochondrial energy sources are impaired, the pump falters, the gradients bleed, and cellular functioning becomes disturbed.  There is extensive recent data of reduced mitochondrial energy generation in bipolar disorder [6, 7].  Whether acting at the level of the membrane or the actual energy supply, lithium has been found to lower intracellular concentrations of sodium and calcium in overactive neurons.
  1. Neurotransmitter systems. Here we return to the catecholamine and other chemical messenger molecules that enable communication between nerve cells.  Lithium acutely increases serotonergic transmission which may mediate its antidepressant effect.  It has an interesting, slower and dual effect, opposing excess excitatory glutamatergic signaling and stimulating inhibitory GABA-ergic activity.  As such, it tends to modulate extremes bringing the overall level of central nervous system buzz towards a balanced middle.
  1. Second messenger systems. Time to extend our metaphor.  Think of the monoamines as letters that are sent between different households.  Think of a large estate, a mansion, with staff of many kinds – landscaping, housekeeping, security – each with their own policies and procedures.  The letters carry instructions on how the estate should modify its existing routines:  Mow the grass less often, use a different cleaner for the bathrooms, add alarms on all the windows, etc….  These internal household procedures are equivalent to a cell’s second messenger systems:  a vast set of inner regulatory pathways that govern what and how a cell carries out its mission.  For example, one pathway might be analogous to the budgeting practices which determines how much money is spent, levels of discretionary income, and rules for saving.  Other pathways might reflect housekeeping, repair and caretaking routines.  Each has significant effects on how the entity functions.  In cellular terms, these inner workings involve the operation of dozens of complicated, intersecting enzymes and protein synthesis sequences.   In this metaphor, the letter carries the information or the signal to the mansion; this is cell signaling.  The execution of the letter’s instructions, which involves change in the facility procedures, is referred to as signal transduction.  The image below conveys the dizzying complexity of this second messenger system trafficking.  Hold on to a chair as you look at this.



The major intracellular pathways that are modified by lithium involve protein kinase A, adenylate cyclase (AC), glycogen synthase kinase 3 beta (GSK-3B), protein kinase C, and the phosphoinositide (PI) cycle.  Each have a cascade of downstream effects on cellular functioning, ranging from modulating neuronal excitability, regulating neurotransmission, and increasing mitochondrial energy generation, to more structural effects on brain growth which I’ll get to shortly.   Of these identified pathways, over-activity of GSK-3B and the PI cycle are most strongly linked to the pathophysiology of manic depression.   Lithium has been found to inhibit the activity of the former and dampen the PI cycle resulting in reduced myoinositol production.  These second messenger system dysfunctions and their correction by lithium are two of the most empirically supported mechanisms of action of this drug.


  1. Transcription factors. Let’s return to the mansion metaphor.   Each estate has its own charter that determines what it is and what it does.  One might be a primary dwelling for a family, another might focus on raising horses, and a third might function as a museum.  For a cell, this basic charter resides in, and is determined by, its DNA.  Transcription factors are intracellular molecules that determine how these genetic policies will be carried out.  When neurotransmitter signaling results in a change in a cell’s basic genetic programming – move the family out and start renting the place, change the landscaping and let the grounds grow wild, close the museum wing –  this is the ultimate level of impact.  The major transcription factor implicated in bipolar disorder is CREB:  cyclic AMP response element-binding protein.  Activation of CREB results in increases in neurotrophic factors which support brain growth and reductions in apoptotic molecules which cause cell death.  Here too, lithium has been shown to modify CREB transcription activity.


  1. Mitochondrial energy production. This point deserves more emphasis.  As mentioned, the mitochondria are intracellular structures that function as energy powerhouses.  They do this through a complicated chemical process called cellular respiration that ultimately produces the compound that fuels most cellular functions, ATP.  This energy pathway also clears the cell of harmful molecules, free radicals, that cause damage and death to neurons.  In estate terms, the mitochondria are the generators that power the operation.  When they falter, the electricity shorts out, the lights flicker, and the heat dwindles.  If persistent and severe, the entire estate can become rundown and decrepit.  Numerous and increasing studies, using a variety of methodologies, are documenting mitochondrial dysfunction in manic depression [6,7].  Related evidence indicates that lithium, acting through several pathways, corrects this impairment, thereby restoring energy production and all it powers –  neurotransmission, maintenance of ion gradients, and regulation of cell excitability –   along with reducing free radicals which results in enhanced cell survival.


  1. So far, this review has emphasized cellular functioning and its residential analogue, estate operations:  what is done, how, regulatory actions, and procedural routines.  The other major dimension of the central nervous system is the structure itself:  the number and size of neurons, the number and size of glia (a type of matrix support cell in the brain), their respective growth, retraction and resilience to withstand injury; and how the growth of these individual cellular units affects the structure and functioning of larger brain areas and brain circuits.  This corresponds to the size and health of our home compound, it’s grounds, and interactions with other estates.  What’s this got to do with bipolar disorder?  Answer:  brain structure is diminished in this disease.  Specifically, the size of the brain areas that mediate emotional experience and emotional regulation have been found to be smaller in those patients with this illness.  This includes the prefrontal cortex, the amygdala, the hippocampus, the striatum, and the anterior cingulate.  There’s a silver lining though:  lithium reverses this shrinkage; it grows brain tissue.  Neuroimaging and post-mortem studies document this remarkable fact.  Through a combination of growth promotion (neuroproliferation) and neuroprotective actions, lithium counteracts the neurodegenerative changes associated with this disease.  In their 2009 paper, Machavado-Viera and colleagues from the NIMH assert that this is the final, convergent pathway for lithium’s therapeutic action.  All operations ultimately support the structural integrity of the estate.  If the estate functions well, it grows, develops, spreads and establishes more connections with other homes.  If various communication and procedural pathways are compromised, the residence withers and may shut down altogether.   How does lithium accomplish this astonishing feat?


Several of the second messenger systems described earlier generate molecules that have either neurotrophic or neurodegenerative effects.  The primary neurotrophic molecules are brain derived neurotrophic factor (BDNF) and the anti-apoptotic protein, B-cell lymphoma 2 (Bcl-2).  The structural bad guys are GSK-3B, the free-radicals mentioned earlier, intracellular calcium, and a host of other culprits.


  1. “Excitable Boy, they all said, Excitable Boy” (Zevon, W; 1978). Using an entirely new experimental procedure, Mertens and colleagues published a report in Nature in 2015 that attempted to clarify the mechanism of action of lithium.  They removed easily-obtainable connective tissue fibroblast cells from 6 patients with bipolar disorder and 4 unaffected individuals.  They then de-differentiated these cells, wiping their genetic operating systems clean, and bringing them back to their original, pluripotent state; the state which all cells are in before they got slotted and programmed into their various specialist roles.   From there the cells were then induced or reprogrammed to become hippocampal neurons, one of the cell types that has regularly shown dysfunction in bipolar disorder.  This induced pluripotent stem cell technology enabled them to compare the hippocampal neurons derived from bipolar patients to those from a non-ill comparison group.  Here’s what they found.  First, the bipolar hippocampal neurons showed abnormalities in the expression of a number of genes, but especially the genes regulating mitochondrial size and function.  Second, the bipolar hippocampal neurons were hyperexcitable; they responded to signals at lower thresholds and with greater chemo-electric responses than control neurons.  Last, and most mind-blowing, this hyperexcitability was controlled by lithium, but only among the hippocampal neurons derived from the bipolar patients who were clinical responders to lithium.   Is hippocampal neuron excitability an endophenotype – the molecular expression – of bipolar disorder?


Where does this review leave us?  We have described a multifaceted picture of bipolar pathophysiology, circa 2016, that includes cell membrane abnormalities, disturbed ion gradients, impaired neurotransmission, aberrant second messenger system signal transduction pathways, transcription factor changes, corrupted mitochondrial function and energy dynamics, neuronal loss and shrinkage in emotion-mediating pathways, and hippocampal neuronal excitability.   (One area which we did not cover was inflammatory changes that are also receiving great attention in our field).  In this review, we’ve tried to use the metaphor of a residential estate – a vast country home with numerous departments involving housekeeping, landscaping, repair, energy generation, and communications – to enable imagination of this complex cellular entity.  No easy task.  If we stick with this metaphor, what does it suggest for our initial question about what we tell our patients when they ask how lithium works?

Help wanted:  Seeking a handy, versatile molecule that can help with all aspects of the management of our large, complex residential estate.  These responsibilities will include border patrol, gatekeeping, oversight of all household departments and procedures, communication systems, landscaping, facilities management, power generation, security and protective services, and the repair and expansion of the grounds.  This is an enormous, multifaceted job but one which is eminently manageable by the right party.  In short, we’re looking for that special, unique molecule with whom we’ll have a perfect chemistry.

John Gottlieb, M.D.


  1. Coppen, A., The biochemistry of affective disorders. Br J Psychiatry, 1967. 113(504): p. 1237-64.
  2. Alda, M., Lithium in the treatment of bipolar disorder: pharmacology and pharmacogenetics. Mol Psychiatry, 2015. 20(6): p. 661-70.
  3. Malhi, G.S., et al., Potential mechanisms of action of lithium in bipolar disorder. Current understanding. CNS Drugs, 2013. 27(2): p. 135-53.
  4. Machado-Vieira, R., H.K. Manji, and C.A. Zarate, The role of lithium in the treatment of bipolar disorder: convergent evidence for neurotrophic effects as a unifying hypothesis. Bipolar disorders, 2009. 11(Suppl 2): p. 92-109.
  5. Mertens, J., et al., Differential responses to lithium in hyperexcitable neurons from patients with bipolar disorder. Nature, 2015. 527(7576): p. 95-9.
  6. Konradi, C., et al., Molecular evidence for mitochondrial dysfunction in bipolar disorder.[Erratum appears in Arch Gen Psychiatry. 2004 Jun;61(6):538]. Archives of General Psychiatry, 2004. 61(3): p. 300-8.
  7. Tang, V. and J.-F. Wang, Mitochondrial Dysfunction and Oxidative Stress in Bipolar Disorder, in Systems Biology of Free Radicals and Antioxidants, I. Laher, Editor. 2014, Springer Berlin Heidelberg: Berlin, Heidelberg. p. 2411-2429.





Affective Temperaments Part II: Approach, Avoidance and the Neurocircuitry of Personality

In my previous blog, I described “temperament” as one’s constitutionally determined way of responding emotionally to the world and introduced Akiskal’s five affective temperaments:  Depressive, Hyperthymic, Cyclothymic, Irritable, and Anxious (1).  Akiskal’s conceptualization of temperament begins with endophenotypes; that is, the outward expression of a gene or trait.   These descriptive endophenotypes provide the scaffolding from which one learns about the underlying processes involved in their formation and evolution.   This approach to studying personality and temperament can best be described as “top-down” – start with what you observe and go from there, from macro to micro.   Another example of the “top-down” view includes the work of Eysenck, who described three “dimensions” of human personality:  Neuroticism (N), Psychoticism(P), and Extroversion (E) (2).   Still another example is that of Cloninger, who described temperament using the following terms: novelty seeking, harm avoidance, reward dependence, persistence, self-directedness, cooperativeness, and self-transcendence (3).

While the “top-down” approach plays an important role in identifying and describing the wide range of temperament, others have used brain circuitry as their starting point.  This would be a “bottom-up” approach, in which what is learned at the level of neurons, neurotransmitters and neural networks lays the groundwork for observed endophenotypes.  I will now attempt to summarize what we know to date about this circuitry, as well as how it can help us understand approach/avoidance patterns of behavior that underlie personality traits as well as anxiety, depression and mania.

In 1991, Jeffrey Gray, a British psychologist, proposed that temperament is formed on the basis of three neurologically based systems: the Behavioral Approach System (BAS), the Behavioral Inhibition System (BIS) and the Fight/Flight System (FFLS) (4). This later became known as the Reinforcement Sensitivity Theory (RST), based on the observation that reinforcement is a prominent feature of the BAS, BIS and FFLS.  Briefly, reinforcers are responses from the environment that increase the probability of a behavior being repeated. Reinforcers can be positive (reward) or negative (removal of an adverse stimulus)(5).    For example, Gray defined the Behavioral Approach System (BAS) as activated by signals of reward and/or removal of punishment.  Both the Behavioral Inhibition System (BIS) and Fight/Flight System (FFLS) would be activated by nonreward and/or punishment; however, in the BIS, all behavior would be inhibited, while in the FFLS, the subject would engage in active escape (flight) or defensive aggression (fight).(4)

Assuming that different neurocircuitry is involved for each component of the Reinforcement Sensitivity Theory (RST) – for the BAS, BIS and FFLS – how exactly does this translate into the formation of temperament?  According to Gray, “Temperament reflects individual differences in predispositions towards particular types of emotions” (4).  One way of interpreting this idea is an individual comes into the world with his or her own pre-programmed way of behaving in response to the environment.  This “programming” sets the individual up to respond with specific behaviors (approach, avoidance, or fight/flight) to “reinforcing” environmental stimuli.   Repeated interactions with reinforcers  establishes the outward manifestation (endophenotype) of temperament.  For example, a person who has several close relatives with severe anxiety may come into the world “primed” for fearful reactions to the environment.   In this person, the Fight or Flight System (FFLS) might be triggered by seemingly harmless stimuli that are associated with what caused the initial fear response (flight).  Or the Behavioral Inhibition System (BIS) may be overactive, causing the person to feel paralyzed in multiple situations.  The pattern becomes “reinforced,” over time, solidifying his or her affective temperament.

Gray’s work has been expanded and supported by Depue and Collins, who in 1999 proposed a neurobiological model of incentive reward and extraversion (6).  They began with Eysenck’s concept of extraversion, focusing on components of interpersonal engagement and impulsivity, and created a model of “positive incentive motivation.  ” This model is thought to be triggered by signals of reward and is based on the Behavioral Approach System (BAS).   Communication throughout the reward circuit (which includes such brain structures as the nucleus accumbens, ventral tegmental nucleus, medial orbital prefrontal cortex, hippocampus and amygdala) is maintained by the neurotransmitter dopamine.  In this circuit, there is an emphasis on “appetitive stimuli” that creates a situation of “incentive motivation” (i.e. the feeling of “I’ve got to have this now!!!”) and results in approach towards the desired object.   Individual differences in dopamine functioning are thought to underlie one’s sensitivity to reward and triggering of the BAS.  One can see how dysregulation of the BAS can get a person into trouble.  Indeed, BAS sensitivity has been implicated in mania, substance abuse, compulsive gambling and ADHD (7).

The idea of Gray’s Reinforcement Sensitivity Theory (RST) being applied to modern research on mood disorders offers intriguing possibilities according to Gonen et. al.   Based on neuroimaging and other data, they propose an “integrative model” of bipolar disorder in which “dysregulated interactions of both punishment and reward related processes… account for the psychological and neural abnormalities observed in (bipolar disorder) (8).”  Interactions between motivational systems (BAS and BIS) determine mood state – mania or depression.   For example, BAS over-activation (enhanced reward sensitivity and approach behavior) with BIS under-activation (reduced punishment sensitivity and avoidance) would result in a manic state.  Conversely, BIS over-activation (enhanced punishment sensitivity and avoidance) with BAS underactivation (reduced reward-driven approach behavior) would result in depression (8).

Gray’s Reinforcement Sensitivity Theory (RST) proposes a model of temperament that starts at the level of neural circuitry.  These circuits reflect an inborn susceptibility to signals of reward (high BAS sensitivity) or punishment (high BIS sensitivity).  Interactions between the individual and the environment reinforce such sensitivities and result in longstanding patterns of behavior.   Depue and Collins use the RST as a starting point in describing a model of positive incentive motivation fueled by dopamine.  Finally, Gonen et. al. apply the RST to their integrative model of bipolar disorder, which explains mood states in terms of inbalances between BAS and BIS.    While our understanding of these processes is far from complete, each of these contributions offers a means by which we can translate brain circuitry into an outward manifestation of mood or temperament.

Susan Stern, M.D.




  1. Akiskal HS, Mendelowicz MV et al. TEMPS-A: validation of a short version of a self-rated instrument designed to measure variations in temperament.  J Affective Disord 2005; 85: 45-52.
  2. Eysenck HJ. General features of the model. In Eysenck HJ (Ed), A Model for Personality. Berlin: Springer-Verlag, 1981: 1-37.
  3. Cloninger CR. A psychobiological model of temperament and character. Arch Gen Psychiatry 1993; 50: 975-990.
  4. Gray JA. The neuropsychology of temperament. In Strelau J et al (Eds), Explorations in New York: Springer Science and Business Media, 1991: 105-128.
  5. McLeod SA. Skinner – operant conditioning.  From simplypsychology.org/operant-conditioning.html, 2015.
  6. Depue RA and Collins PF. Neurobiology of the structure of personality: dopamine, facilitation of incentive motivation, and extraversion.  Behav Brain Sci  1999; 22: 491-517.
  7. Farmer RF. Temperament, reward and punishment sensitivity, and clinical disorders: implications for behavioral case formulation and therapy.  Intl J Behav Cons and Therapy 2005; 1: 56-76.
  8. Gonen T et al. Moods as ups and downs of the motivation pendulum: revisiting reinforcement sensitivity theory (RST) in bipolar disorder.  Frontiers Behav Neurosci 2014; 8: 1-8.

Affective Temperaments Part I: What Can Personality Traits Teach Us About Mood States?

Emotional expression, or “affect,” covers a range of temporal domains.  There are “emotions,” moment-to-moment fluctuations which, while intensely experienced, come and go within minutes.  When a given emotional state lasts longer – hours, days, or months – it is described as “mood.”  Finally, there is “temperament,” a lifelong emotional disposition considered to be part of one’s constitutional makeup (1).  When temperament manifests as “affective” – that is to say, appears as a similar but less severe variant of a disordered mood state – things start to get interesting.

How common are affective temperaments?    Affective temperaments are thought to be present in up to 20% of the general population (2).   According to one model, there are five types of affective temperaments– Depressive, Hyperthymic, Cyclothymic, Anxious, and Irritable.   In both research and clinical settings, the presence of one of more of these affective temperaments can be established by administration of the TEMPS-A questionnaire (Temperament Scale of Memphis, Pisa, Paris and San Diego)(3).

The relationship between affective temperaments and affective illness appears to be complex, with some researchers indicating that affective temperaments represent a “latent stage” of illness (2) while others place affective temperaments on one end of a “spectrum” that leads to more severe forms of illness (4).  I will now consider three studies which together help to elucidate the relationship between affective temperaments and affective illness.

In 1992, researchers at the University of Pisa and University of Tennessee analyzed data from 538 patients presenting with a major depressive episode.   Unexpectedly, they found that patients with unipolar depression superimposed on hyperthymic temperament presented a unique subcategory (UP-HT).  Hyperthymic temperament is described as trait exuberance, cheerfulness, talkativeness and extraversion (5).  The UP-HT subgroup had some features comparable to unipolar depression (such as age of onset, presence of melancholia) but others more similar to patients with bipolar disorder (such as equal male/female sex ratio, higher rate of first degree relatives with bipolar disorder).  These findings supported the “validity of using hyperthymia as a temperamental indicator of bipolarity in patients suffering from (major depressive episode)” (5).

The 1992 paper is important because it established that the course of affective illness, in this case depression, could be modified by underlying affective temperament.  This UP-HT variant, which has genetic similarities to bipolar disorder (similar sex distribution and family history) starts to look more like it should be on a spectrum of bipolar disorder; indeed, in later papers Akiskal referred to UP-HT as “Bipolar Type IV” (4).   In this case, consideration of one’s underlying hyperthymic affective temperament in the setting of depression would be helpful in guiding the clinician towards appropriate treatment (i.e. monitoring for antidepressant-induced hypomania, considering mood stabilizer treatment earlier).

In 2003, Akiskal, Hantouche and Allilaire published a study focusing on bipolar II disorder, with (n=74) and without (n=120) cyclothymic temperament (CT) (6).  Cyclothymic temperament has been described as a combination of biphasic, abrupt mood swings along with at least four of the following: alternations between lethargy and eutonia, low self-confidence and overconfidence, decreased verbal output and talkativeness, mental confusion and sharpened/creative thinking, tearfulness and jocularity, and introverted self-absorption and uninhibited people-seeking (4).   The group found that patients with bipolar II disorder with CT had a younger age of onset of illness, more episodes of depression and delayed recognition/diagnosis of bipolar disorder.  Most significantly, the cyclothymic bipolar II group scored significantly higher on “irritable risk taking” than “classic driven-euphoric” items of hypomania.  This led to the establishment of “Bipolar II 1/2” (Bipolar II with cyclothymic temperament) – a more “unstable” and “dark” variant of bipolar II disorder (6).  Early recognition of “dark” bipolar II disorder is key in order to closely monitor the individual, as the combination of irritability, depression, and impulsivity can have dangerous consequences.

While the above two studies represent some of the best and most interesting work in the field of affective temperaments, questions remain.   Specifically, larger studies have been lacking that compare the frequency of particular affective temperaments in clinical and non-clinical populations.  More recently, however, a meta-analysis of 26 studies has been published comparing TEMPS scores across mood disorder patients, their first-degree relatives, healthy controls, and other psychiatric disorders (7).

The researchers found that patients with bipolar disorder (BD) had significantly higher cyclothymic (P<0.001), hyperthymic (P<0.001) and irritable (P<0.001) TEMPS scores compared to patients with major depressive disorder (MDD).  Depressive and anxious TEMPS scores were not different between the two groups.   When comparing bipolar disorder type I (BP-I) with bipolar disorder type II (BP-II) patients, depressive TEMPS scores were lower in BP-I compared with BP-II (P=0.002).  This latter finding could lend validity to the clinical observation that BP-II patients spend much more time in a depressive state than in a hypomanic one.  In comparing bipolar disorder (BD) to healthy controls (HC), bipolar patients had significantly higher TEMPS scores for cyclothymic, depressive, irritable, and anxious temperaments (P<0.001) with hyperthymic TEMPS scores being higher in the HC group than BD group (P<0.001).  Findings were similar when comparing MDD to HC, indicating that having hyperthymic temperament is likely a protective factor for both unipolar and bipolar disorders.

Comparisons of TEMPS scores for bipolar patients in comparison with first-degree BD relatives and healthy controls (HC) are of interest as well.   Cyclothymic (P<0.001), irritable (P=0.001) and anxious (P=0.03)TEMPS scores were significantly higher in the BD group compared with BD relatives.  In comparing first-degree BD relatives with HC, cyclothymic (P=0.007), irritable (P<0.001) and anxious (P=0.01)TEMPS scores were significantly higher in BD relatives than in HCs.

The results of the meta-analysis help to validate the idea of mood disorders as being on a continuum, with TEMPS scores for cyclothymic and irritable temperaments increasing from healthy controls (HC) through major depressive disorder (MDD) to bipolar disorder (BD).  TEMPS scores for hyperthymic temperament increased from MDD through BD to HC.  For cyclothymic, irritable and anxious temperaments, TEMPS scores increased from HC through BD relatives to BD (7).

To summarize, data from the past two decades indicate that affective temperaments and mood disorders are closely linked.    Akiskal et. al. used their findings to establish the “soft bipolar spectrum”(4) model of illness, which incorporates underlying cyclothymic temperament in combination with recurrent depressive illness.   Cyclothymic temperament was found to be a risk factor for the development of bipolar disorder as well as a complicating factor.   Cyclothymic temperament in combination with bipolar II portends a “dark” variant of hypomania characterized by irritability impulsivity and high-risk behavior.   More recently, a meta-analysis has reinforced the associations between affective temperaments, unipolar vs. bipolar disorder, first-degree relatives and healthy controls.   Important questions do remain about the precise nature of these associations.  For example, what is the % risk that a patient with a given affective temperament will develop bipolar disorder, and by what process?  Only longitudinal, prospective studies following individuals with affective temperaments could begin to answer this question.  The complex interactions between the “trait” characteristics of affective temperaments and “state” mood disorders are fascinating and have important implications for both diagnosis and treatment.

Susan Stern, M.D.



  1. Goodwin FK and Jamison K. Manic-depressive illness, 2nd  Oxford University Press, 2007: p. 324, 609.
  2. Gonda X and Vasquez GH.  Theoretical and clinical overview of affective temperaments in mood disorders.   Psicodebate 2014; 14: 39-58.
  3. Akiskal HS, Mendlowicz MV et al. TEMPS-A: validation of a short version of a self-rated instrument designed to measure variations in temperament.  J Affective Disord 2005; 85: 45-52.
  4. Perugi G and Akiskal HS. The soft bipolar spectrum redefined: focus on the cyclothymic, anxious-sensitive, impulse-dyscontrol, and binge-eating connection in bipolar II and related conditions.  Psychiar Clin N Am 2002; 25: 713-737.
  5. Cassano GB, Akiskal HS et al. The importance of measures of affective temperaments in genetic studies of mood disorders.  J Psychiat Res 1992; 26: 257-268.
  6. Akiskal HS, Hantouche EG and Allilaire JF. Bipolar II with and without cyclothymic temperament: “dark” and “sunny” expressions of soft bipolarity.  J Affective Disord 2003; 73: 49-57.
  7. Solmi M, Zaninotto L et al.  A comparative meta-analysis of TEMPS scores across mood disorder patients, their first-degree relatives, health controls, and other psychiatric disorders.  J Affective Disord 2016; 196: 32-46.


Can Anti-inflammatory Drugs Treat Depression? Some Promising New Evidence But Not Yet Enough

The connection between physical ailments and mood is not a new one. It is well-established that medical illness, ranging from infections to cardiovascular disease, can result in increased symptoms of depression, while depression can predispose people to become physically sick more often.  Now, a growing body of evidence shows that depression and physical illness have something important in common: inflammation.

In theory, this connection hints that medications such as NSAIDS (non-steroidal anti-inflammatory drugs) might be helpful in the treatment of depression. This article will attempt to summarize how inflammation – a generalized process affecting the whole body – can also enter into the brain and affect mood, and how anti-inflammatory medications such as NSAIDs have the potential to put a stop to this irritating scenario.

As an example of the growing evidence base on this very topic, consider this new study by a team of Danish researchers (Kohler et al, 2014). This group performed a meta-analysis (a statistical review that allows for the pooling of data) reviewing “the antidepressant and possible adverse effects of anti-inflammatory interventions” (1). Fourteen trials (6262 participants) were included in the study, ten of which evaluated the use of NSAIDs (n=4258).  The group performed a “pooled effect estimate” that suggested that anti-inflammatory treatment, particularly the NSAID celecoxib, decreases depressive symptoms without increased risks of adverse effects.    Despite the caveat that the studies included a wide range of patients on a variety of treatments (including in some cases studies combining NSAIDs and antidepressants (2)), they state that “it is possible that specific subgroups would benefit more from anti-inflammatory intervention, such as patients with low-grade inflammation or co-morbid inflammatory diseases” (1).

This study is worth noting because it captures a trend across multiple studies: anti-inflammatory medications seem to have a positive effect on depression. But how do anti-inflammatory drugs such as celecoxib exert an anti-depressant effect? In order to answer that question, it is important to understand the process of inflammation both peripherally (in the body) and centrally (in the brain).

Our immune systems are on the lookout for evidence of invaders, both real and imagined. In the presence of infection, cellular damage, or stress, the inflammatory response may be “appropriate and necessary to maintain homeostasis in the body.” However, the inflammatory response may be “inappropriate, pathological and damaging when it is reacting out of proportion to a given stimuli or reacting to the wrong stimuli” (3).   Whatever the cause, inflammation can induce “sickness behavior,” which includes many of symptoms of depression: low mood, lethargy, decreased appetite, and diminished interest in many activities (3).

The presence of concurrent peripheral inflammation and symptoms of depression (as manifested by “sickness behavior,” for example), indicates involvement ofboth systemic inflammation and central neuro-inflammation.  In order for this to happen, mediators of peripheral inflammation, such as pro-inflammatory cytokines (small proteins that participate in local and systemic inflammatory effects (3)), must be able to infiltrate the central nervous system (CNS).  Pro-inflammatory cytokines include interleukins (including IL-1 beta and IL-6) and tumor necrosis factor (TNF-alpha). Cytokines play an important role in amplifying the effects of inflammation throughout the body.

The “cytokine hypothesis” offers a model by which cytokines travel to the CNS and cause changes in brain functioning that contribute to depressive symptoms. Several mechanisms are proposed by which cytokines get into the CNS, including hitching a ride on nerves and blood vessels leading to the brain. Once in the CNS, cytokines are thought to break down and/or decrease the production of serotonin and to contribute to a “chronic stress state” marked by dysfunction of the HPA (hypothalamic-pituitary axis) and chronically elevated cortisol levels (4). Microglia, specialized cells that function as “the resident immune sentinels” of the brain, are activated by the presence of inflammatory activity and contribute to the production of even more inflammatory cytokines. Left unchecked, chronic inflammation in the CNS can destroy functional neuronal pathways by inciting those same microglia to engage in abnormal pruning of synapses, resulting in maladaptive changes to brain structure and function (3,4).

So where do NSAIDs come in? Pro-inflammatory cytokines in the CNS stimulate the breakdown of cell membranes. Cyclo-oxygenases COX-1 and COX-2) are enzymes that use these breakdown products to produce prostaglandins, leukotrienes and thromboxanes (5), which further amplify the inflammatory reaction. NSAIDs arrest this process by stopping the action of one or more of the cyclo-oxygenases. For example, celecoxib inhibits COX-2. In animal models, celecoxib has been shown to decrease pro-inflammatory cytokine production as well as to increase serotonin production (6).   All of which is considered a good thing for not only decreasing inflammation, but also for addressing concurrent symptoms of depression.

The data from the Danish group is promising, yet there are many unanswered questions. Which NSAIDs are safest and most effective, and in which patients? Should they be used alone or together with conventional antidepressants? How long should treatment last, especially since the majority of the trials were only 6 weeks in duration?

In reality, we are far from ready to make evidence-based recommendations for initiating anti-inflammatory medications for the treatment of depression.   In the absence of more data, anti-inflammatory medications should be reserved solely for the treatment of underlying inflammatory-based medical conditions.   But it is intriguing to consider the possibility that for such patients, their NSAID is addressing both physical and emotional symptoms.

Susan Stern, M.D.



  1. Kohler O et al. Effect of anti-inflammatory treatment on depression, depressive symptoms, and adverse effects: a systematic review and meta-analysis of randomized clinical trials. JAMA Psychiatry. Published online October 15, 2014.
  2. Fond G et al. Effectiveness and tolerance of anti-inflammatory drugs’ add-on therapy in major mental disorders: a systematic qualitative review. Acta Psychiatrica Scandinavica 2014; 129: 163-170.
  3. Rosenblat JD et al. Inflamed moods: a review of the interaction between inflammation and mood disorders. Prog Neuro-Psychopharmacol Biol Psychiatry. Published online January 13, 2014.
  4. Jones KA and Thomsen C. The role of the innate immune system in psychiatric disorders. Molecular and Cellular Neurosci 2013; 53: 52-62.
  5. Farooqui AA et al. Modulation of inflammation in brain: a matter of fat. J Neurochemistry 2007; 101: 577-599.
  6. Muller N. The role of anti-inflammatory treatment in psychiatric disorders. Psychiatria Danubina 2013; 25: 292-298.




Understanding Premenstrual Dysphoric Disorder (PMDD)

Imagine a mood disorder in which intense mood swings come predictably every month, wreaking havoc on any idea of calm normalcy. Depressed mood, lethargy, decreased interest and hopelessness occur along with marked irritability, anger, agitation and insomnia.   One has the sense of being overwhelmed and “out of control.” Arguments and heightened tearfulness ensue. Then everything returns to normal for the next week or two, only to be turned upside down by the dreaded monthly roller-coaster.

The intensity of what I have described sounds similar to bipolar disorder in its dramatic presentation. However, it is actually a fairly typical description of Premenstrual Dysphoric Disorder (PMDD), which has recently been recognized as a mood disorder in the DSM-V (1).   The criteria for PMDD is strict in terms of timing, duration, and degree of functional impairment. Symptoms begin during the luteal phase (days 15-28 of the menstrual cycle) and peak around the time of onset of menses. While symptoms can linger into the first few days of menses, there must be a symptom-free period in the follicular phase after the menstrual period begins. The diagnosis needs to be confirmed by prospective charting of at least two symptomatic cycles, and cannot be an exacerbation of symptoms of another disorder (such as depression or bipolar disorder), though PMDD can co-occur with these disorders (1).

PMDD is considered less common than PMS, with an estimated prevalence rate of 3-8% (2) compared with 75% of women for PMS (2,3).   The hormonal fluctuations of the menstrual cycle are themselves not the cause of the mood dysregulation. Rather, certain subpopulations of women have been observed to develop sensitivity to the normal hormonal fluctuations of the menstrual cycle (5). Clinically, we have observed such women to include those who have a personal or family history of mood disorders, or those who seem to have a heightened response to environmental stress.

Multiple mechanisms have been proposed for PMDD. Estrogen (primarily E2, or Estradiol), can affect the serotonin system in ways that are similar to SSRIs. In both rodent and human studies, E2 receptors have been found to be plentiful in the hippocampus and amygdala and modulate the affective response to stress (6). In the normal luteal phase in humans, levels of E2 drop dramatically, which, in some women, could trigger a heightened negative response to stress through several proposed mechansisms. For example, E2 normally decreases MAO (monoamine oxidase) activity, which increases the availability of serotonin (5-HT), dopamine (DA) and norepinephrine (NA). A decrease in E2 could thus trigger a depressive reaction by increasing the breakdown of these neurotransmitters by MAO (2). Attention has also been focused on the role of E2 in the formation of dendritic spines on pyramidal cells in the hippocampus and prefrontal cortex of the brain. Dendritic spine formation is considered to enhance both mood and cognition, with E2 and brain-derived neurotrophic factor (BDNF) working in concert (7). During the follicular phase of the menstrual cycle, when E2 is rising, spines form. During the late luteal phase, however, with E2 falling, the spines become dismantled, resulting in depressed mood and poor concentration (8).

The above mechanism of dendritic spine deterioration is also thought to be promulgated by progesterone, which, like estrogen, rises but then falls during the luteal phase of the menstrual cycle (8). Progesterone is also thought to increase MAO activity, decreasing neurotransmitter activity by causing their breakdown (2). Perhaps for this reason, progesterone has been thought of as a “depressogenic” hormone. However, recent research has highlighted another role for progesterone, in the form of its metabolite, the “neurosteroid” allopregnanolone (Allo-P) (9-11). Neurosteroids are endogenous steroids synthesized in the brain and nervous system from cholesterol that are potent modulators of the two major neurotransmitter systems that govern CNS activity – glutamate, the major excitatory neurotransmitter, and GABA (gamma-aminobutyric acid), the major inhibitory neurotransmitter (10). Too little GABA, too much glutamate (the main “excitatory” neurotransmitter) is considered to be a possible mechanism for depression, anxiety, mania and other disorders. In general, Allo-P is thought to be a GABAa receptor enhancer (10).

The drop in progesterone during the late luteal phase of the menstrual cycle causes rapid withdrawal from Allo-P and may be responsible for PMDD symptoms (11). Recent animal studies have shown that low-dose SSRIs such as fluoxetine given during the late luteal phase can stimulate the production of Allo-P, resulting in a rapid improvement in mood symptoms (over hours to days). This is via a separate mechanism from the slower process of selective serotonin reuptake inhibition, in which the patient may not notice a clinical effect for a longer period (at least two weeks). Indeed, the author proposes calling fluoxetine a “selective brain steroidogenic stimulant” (11). The mechanism by which fluoxetine rapidly generates Allo-P production is unclear, but it supports clinical observations of improvement in PMDD symptoms in some patients with intermittent, luteal-phase only dosing of SSRIs (3).

So Allo-P is generally a good thing, correct? Here we run up against seemingly conflicting data – for 3-8% of menstruating women (interestingly, the same prevalence of women who have PMDD), Allo-P causes “paradoxical” effects on the GABAa receptor system (4). The Allo-P-induced negative mood symptoms are dependent on how much Allo-P is present – very low and very high concentrations have less of an effect on mood. However, during endogenous luteal phase levels, negative mood occurs. Women with PMDD are thought to have a “supersensitive” GABAa receptor in which Allo-P actually changes the configuration of the receptor so that it no longer functions as an inhibitory receptor and instead causes “paradoxical” heightened anxiety, depression, and irritability during the luteal phase of the menstrual cycle (4).

So what are the clinical implications of these findings? It appears that for women who have true PMDD, a “less is more” approach applies to treatment, especially use of SSRIs. The trick appears to be how to find and hit the “sweet spot” of just enough Allo-P produced to keep the GABAa receptor working the way it’s supposed to. The desired result – relief from the emotional vicissitudes of PMDD – may be a few steps closer. However, more research must be done to fully clarify the pathophysiology of this elusive illness. Doing so may also shed some light on why some women, but not others, are susceptible to mood shifts in response to the wide ranges of hormone levels found in a typical reproductive lifespan.



  1. Diagnostic and Statistical Manual of Mental Disorders, 5th ed.  Arlington, VA: American Psychiatric Association, 2013.
  2. Spinelli M. Neuroendocrine effects on mood. Reviews in Endocrine and Metabolic Disorders 2005; 6: 109-115.
  3. Steiner M et. al. Expert guidelines for the treatment of severe PMS, PMDD, and comorbidities: the role of SSRIs. Journal of Women’s Health 2006; 15: 57-69.
  4. Backstrom T et. al. Allopregnanolone and mood disorders. Progress in Neurobiology (online) July 5, 2013: 1-7.
  5. Joffe H and Cohen LS. Estrogen, serotonin, and mood disturbance: where is the therapeutic bridge? Biological Psychiatry 1998; 44: 798-811.
  6. Ter Horst GJ. Estrogen in the Limbic System. Vitamins and hormones 2010; 82, 319-338.
  7. Luine V and Frankfurt M. Interactions between estradiol, BNDF and dendritic spines in promoting memory. Neuroscience 2013; 239: 34-45.
  8. Stahl S. Understanding and managing the pieces of major depressive disorder. Carlsbad, CA: Neuroscience Education Institute, 2009, 103-104.
  9. Zorumski CF and Mennerick S. Neurosteroids as therapeutic leads in psychiatry. JAMA Psychiatry 2013; 70: 659-660.
  10. Zorumski et. al. Neurosteroids, stress and depression: potential therapeutic opportunities. Neuroscience and biobehavioral reviews 2013; 37: 109-122.
  11. Lovick TL. SSRIs and the female brain – potential for utilizing steroid-stimulating properties to treat menstrual cycle-linked dysphorias. Journal of Psychopharmacology (online) May 23, 2013: 1-6.


Many of us of a certain age will never forget the sage, one-word piece of occupational advice given to Benjamin Braddock in the movie, The Graduate:   “Plastics.”  In the early 1960’s, plastics were thought to be the next big area of economic growth.  Benjamin, as we all remember, takes in this recommendation with alienated befuddlement.  However from the vantage point of 21st century psychiatric research, this advice may have been unwittingly prescient.

The plasticity of brain structure, or neuroplasticity, has become a major focus of scientific study over the last 10 to 15 years.  Neuroplasticity refers to the degree to which the organization and function of the brain changes through experience.  This article will briefly review the background of this area, some basic findings, and its relevance for understanding and treating bipolar disorders.

Several factors stimulated research on neuroplasticity.  First, shortcomings of the monoamine theory of depression.  Monoamines, like dopamine, norepinephrine, and serotonin are neurotransmitters involved in mood regulation.  In its most simple form, the monoamine theory postulated that deficiencies in monoamine levels caused depression.  We now know that this model has major limitations.  Foremost among them, the several-week lag period between starting an antidepressant and experiencing symptomatic relief.  Given that antidepressants lift monoamine levels almost immediately, there is obviously another mechanism at work that underlies depressed mood states and their response to medication.  Second, we know that the brains of people with bipolar disorder show structural differences from those without the illness.  Affected individuals exhibit overall reductions in gray matter (the part of the brain composed of nerve cells themselves in contrast to white matter which consists of the nerve axons, the long fibers that snake through the brain connecting one cell to another) , increases in amygdala size, and shrinkage or atrophy of specific areas, such as the hippocampus and the dorsolateral prefrontal cortex.  Notably, it has been found that lithium, valproate (Depakote) and certain antidepressants correct these changes.  The last phenomenon shifting attention towards brain structure has been the increased recognition of cognitive impairment occurring in mood disorders.  Roughly correlating with duration of illness and number of episodes, the occurrence of cognitive impairment – especially when independent of mood states – has suggested a chronic, structural problem with the memory-specific areas of the brain.  Together, these three factors shifted research attention away from purely biochemical, neurotransmitter-based models of mood pathology towards a focus on the underlying wiring and arrangement of neural groups in the brain.  This emphasis on brain structure and its malleability has generated several basic ideas that are guiding investigation in this area.

First, the brain is a dynamic organ that responds to new experiences with microscopic changes in its circuitry and functioning.  Read that last sentence again.  The wiring and connections of neurons actually become modified through experience.  This is the first and most significant finding of this new field.  Based initially on animal studies and more recently through MRI and other neuroimaging techniques in humans, researchers are now able to visualize how neurons respond to new experience.  This includes increases in the size and number of dendrites (the part of the nerve cell specialized to receive input from other nerve cells), weakening or strengthening of synaptic connections, and shrinkage or growth of cell groups.   Using a map analogy, we can think of our genetics as providing the layout of the major thoroughfares and neural pathways of our brain.  It is our individual experiences though, which chart out the streets, alleys, and more personally specific wiring that provide our own distinct psychology and physiology.  Hence, we need to give up the outdated view of the brain as a static given and replace it with something much more like a control center muscle that changes its size and abilities based on experience.  This understanding leads to two new questions:  How quickly and when do these structural changes occur?  Additionally, what types of experience are capable of producing these changes?

The initial perspective on neuroplasticity was that it was a phenomenon confined to childhood.  We know that the brains of all mammals expand and develop throughout infancy.  The earlier view was that this potential for change was limited to a critical period of development and that this window closed as the individual matured into adulthood.  The discovery of new brain growth – neurogenesis – in adult mammals was made only recently.  It has been a game-changer.  With more detailed neuroimaging methods, research has shown that many areas of the brain retain the capacity for growth, regeneration, and modification throughout life.  While diminished from its earlier childhood potential, this lifelong capacity is significant and can be activated by a range of experience and interventions.  What types of experience and interventions act on this neuroplastic foundation?  Short answer:  almost everything.

New experiences, new learning, the acquisition of new skills – all these developments result in, and are mediated by microscopic changes in neural circuitry.  This is the second paradigm shift of this young field.  Not only is the brain a dynamic organ, but it changes on a daily basis in response to everything that happens to us.  The new friend we make, the first job we hold, the unexpected storm that hits, the birth of a child, the loss of a parent, the hilarious joke we hear, the new food we try – the experiences of our lives are constantly being processed by, and changing the cellular architecture of our brains.  This realization sets up the final two ideas of this article:  What does neuroplasticity have to do with bipolar disorder?   And what can be done about it?

Current thought holds that impaired neuroplasticity contributes to the abnormal mood states, impaired cognition, structural changes, and delayed antidepressant response found in bipolar disorders.   In times of health, our brains and minds retain the ability to learn, grow, forget, strengthen, and prioritize certain information.  When depressed though, one might only focus on negative events and memories.  One might forget that his or her depressions are always brief and quickly respond to therapy, and feel instead that their black state will go on forever.  In contrast, in mania it may be difficult to appreciate the risks and difficulties of unrestrained plans and ambitions.  From a neuroplasticity perspective, these moods reflect a condition of cognitive and emotional constraint, and an underlying impairment in learning and brain malleability.  Here too, this theory makes intuitive sense to anyone that has experienced or dealt with the above mood states.   It is often difficult to help a person get perspective (i.e., think differently, remember other times, etc…) on their emotional condition.  Neuroplasticity theory would also suggest that the subtle cognitive impairment and structural changes found in the brains of bipolar individuals reflect the same type of reduced potential for experience-based brain circuit modification.  Through this lens, manic depressive illness is seen as a disorder of structural, cognitive and emotional unresponsiveness, constriction, and shrinkage.  Things get locked in a narrow groove from which one cannot easily escape.  So what’s to be done about this?

Neuroplasticity models suggest a role for therapy, medications, and a host of adjunctive interventions to restore the free-flowing dynamism of a healthy brain.  As a new form of experience, psychotherapy is uniquely suited to challenge, unlock, and expand the cognitive and structural constrictions of extreme mood states.  Psychotherapy is designed to provide new experiences for patients.  Its effectiveness may lie in its ability to do this.  Interestingly, antidepressant medications have also been found to restore the reshaping potential of mammalian brains.  The neuroplasticity model arguesthat it is this mechanism which mediates antidepressant response and accounts for the delay between rising monoamine levels and improvement.  Last, animal research has demonstrated the effects of environmental and behavioral changes on brain structure.  Rats caged in enriched environments show increased brain growth, enhanced neuronal resilience in response to stress and greater dendritic arborization (branching) than their peers housed in standard, barren cages.  This line of research has been extended to exercise, video games, and may also apply to environmental exposure to adequate sunlight, level of socializing, and intellectual stimulation.   This model argues that all effective therapeutic modalities will act by directly or indirectly restoring the mind’s ability to freely process information and the brain’s capacity for experience-guided fine-tuning.

Research on neuroplasticity presents a new way of thinking about the brain and bipolar illness.   It moves us away from a more fixed model of the central nervous system to one that is intrinsically reactive and evolving.  It elaborates on simple ideas about excess or reduced neurotransmitter levels to more complex models that include attention to the scaffolding, layout, and modifiability of the surrounding cellular matrix and nerve pathways.  In so doing, it encourages a wider range of therapeutic interventions for the difficult mood states that characterize this challenging disorder.

Plastics.  Indeed.

In future pieces, we will describe some of important mechanisms and major molecular players involved in mediating neuroplasticity.


  1. Carlson, P. J., J. B. Singh, et al. (2006). Neural circuitry and neuroplasticity in mood disorders: insights for novel therapeutic targets. NeuroRx 3(1): 22-41.
  2. Kapczinski F, Frey BN, Kauer-Sant’Anna M, Grassi-Oliveira R. (2008). Brain-derived neurotrophic factor and neuroplasticity in bipolar disorder. Expert Rev Neurother. 2008 Jul;8(7):1101-13.
  3. Krishnan, V. and E. J. Nestler (2010). Linking molecules to mood: new insight into the biology of depression. American Journal of Psychiatry 167(11): 1305-1320
  4. Bavelier, D., D. M. Levi, et al. (2010). Removing brakes on adult brain plasticity: from molecular to behavioral interventions. Journal of Neuroscience 30(45): 14964-14971.
  5. Castren, E. (2013). Neuronal network plasticity and recovery from depression. JAMA Psychiatry 70(9): 983-989.

Bipolar Disorders and the Case of the Missing Self

Every few years, a new author comes along who is uniquely capable of giving voice to the ineffable aspects of their experience with serious mood problems:  Kay Jamison with her (An) Unquiet Mind, William Styron who perceived Darkness Visible, and Sylvia Plath’s The Bell Jar are the more modern prototypes.    Recently, a freelance journalist, Linda Logan, published a brief piece in the New York Times:  The Problem With How We Treat Bipolar Disorder 1.   This mini-memoir is second-to-none in capturing the roller-coaster ride that is far-too-often the case with this illness.

In prose that is simple, direct, and blunt, Ms. Logan describes her 25 year journey through missed diagnosis, misdiagnosis, and a cavalcade of treatments that alternately helped and worsened her underlying condition.  This is a sobering read; it is not for the faint-of-heart.

Beyond her gift for describing these diagnostic and treatment experiences, Ms. Logan focuses on a grossly neglected aspect of this illness:  what it does to one’s sense of self.  Starting from her early experiences with depression, moving through a kaleidoscope of hypomanic and psychotic states, including the psychological impact of various medications, she describes how the experience of intense mood states and their treatment challenge our most basic knowledge of ourselves.  As clinicians working in this area, we find that questions about self-identity almost inevitably arise in the course of this illness.   When a diagnosis of bipolar disorder is made, for example, people who knew of themselves as outgoing, upbeat, and irreverent are forced to consider whether their fast-paced levity was who they truly were or part of an illness.  Or the person who has spent so long in a depressed state that it comes to define both how they know themselves and how others know them.  This person too will have to consider the same questions if and when their depression is effectively treated.  Illness or identity?   Helping patients with mood disorders address and resolve these almost existential uncertainties is an often necessary part of their therapy.

In her article, Ms. Logan describes her long journey through bipolar disorder and her experiences with losing, questioning, despairing about, and ultimately finding a new sense of herself.  The final version incorporates some of her old and healthy qualities, while acknowledging the ravaging effects of the illness, and ultimately spins out a fresh original that combines bits of old and new, disturbed and undeterred, lost and found.

Bipolar disorder presents many challenges to those it afflicts.  Linda Logan’s writing brightly illuminates one such trial – that of her ‘vanishing self’ – and the torturous path towards recovered and reconstructed identity.   If you or someone you know has this illness, read this work, hold on tight, and gain inspiration from this brave author’s courage.

The Problem With How We Treat Bipolar Disorder.  Linda Logan.  New York Times.  April 26, 2013.  http://www.nytimes.com/2013/04/28/magazine/the-problem-with-how-we-treat-bipolar-disorder.html

Brain Problems in Bipolar Disorder: What We Know and How We Know it, circa 2013.

“Is there something wrong with my brain?”  “Does part of my brain not work correctly?”  Each time we make a diagnosis of bipolar disorder in our practice, these questions inevitably and understandably come up.  People want to know about their illness.   This conversation is often an essential part of the treatment.

In 2012, the journal Bipolar Disorders devoted an entire issue to a review and summary of modern neuroimaging research.  It included articles that presented a model of the signature disturbances found in the bipolar brain1-6.  In this longer-than-average blog piece, we will present these findings.  To make this understandable, we’ll begin with a short introduction about neuroimaging and the concept of functional neuroanatomy.


Over the last 20 years, a new series of neuroimaging techniques has been developed that extended and deepened our appreciation of brain structure and function.  Magnetic resonance imaging (MRI) enabled higher resolution snapshots of smaller brain structures than was possible with older computed tomography (CT) scans.  The advent of functional neuroimaging – with functional MRI, positron emission tomography (PET), and single proton emission computed tomography (SPECT) scans – has allowed us to see which areas of the brain are activated when performing specific mental tasks, such as holding something in memory or looking at an emotionally charged picture.   The new techniques also enable us to look beyond single region assessments.  By comparing co-occurring levels of activation in different brain regions during an emotional or cognitive task, fMRI has been used plot out the functional linkage between these separate regions.  We can see how much two areas work together in performing a particular job, thereby generating maps of functional neural connectivity or circuits.   Diffusion tensor imaging (DTI) complements this by analyzing the integrity of the white matter pathways that connect different brain regions.  The result of these new neuroimaging techniques has been the establishment of an early functional architecture of the human brain.  This architecture transcends the earlier ‘brain region A correlates with mental function B’ approach, replacing it with an appreciation of dynamic neural circuits that travel through and utilize multiple brain areas to support our cognitive and affective needs.   So what has this all got to do with bipolar disorder?

Using these new neuroimaging methods, psychiatric researchers have begun constructing functional models of how the brain processes emotion.  Pathways involved in each of the steps of affective behavior – from the initial recognition of the emotional component of a stimulus (e.g., the anger on the face of a mad gunman), through the first emotional response (e.g., fear), the recruitment of cognitive strategies to deal with the situation (e.g., placating the gunman, negotiating, confrontation etc…), selection of the optimal approach (e.g., avoidance of the threat), and the final mobilization of adaptive behavior (e.g., flight/escape) – are being charted and refined.  The result is an emerging map of what parts of the brain (which circuits) do what and how various cognitive, behavioral, and emotional responses are mediated.  This is functional neuroanatomy:  an engineering-like analysis of how the brain works.

With this growing knowledge of how normal human emotion is processed, the stage was set for the study of bipolar patients, to see how and where their responses and underlying neural circuit activation differs from those without this illness.  Over the last 15 to 20 years, a variety of such ‘compare-and-contrast’ neuroimaging experiments have been performed.  In the following sections, we present a brief overview of these findings.  They contain some of the information that we provide when we attempt to answer our patient’s questions about “What is wrong with my brain?”

Brain Abnormalities Associated with Bipolar Disorder

Finding #1:   The brain systems that support normal human emotional response are the ones involved in bipolar disorder.  This has been both a guiding assumption, and an increasingly validated conclusion of research in this area.   In other words, when researchers first began looking for neural abnormalities in bipolar disorder, they had to choose which, among the vast multitude of brain regions, to focus on.  They assumed that the areas that had been found to be associated with ordinary emotional experience would be the ones most likely to show dysfunction in bipolar disorder.  This assumption looks correct.

Finding #2:  A circuit involving the prefrontal cortex, the amygdala, and their connecting pathways is especially implicated in bipolar disorder. The amygdala is an evolutionarily ancient cluster of neurons deep in the temporal lobe.  This primitive structure is responsible for the most basic aspects of emotional experience that occur in all animal species, from reptiles through humans.  This includes the rapid evaluation of emotional stimuli (e.g., is that lion friendly or looking to eat me?) and the mobilization of an initial affective response (e.g., run!).  The prefrontal cortex (PFC), in contrast, is the evolutionarily most-recent, distinctly human part of the brain that sits on top of lower subcortical areas such as the amygdala.  It controls much of our decision-making, planning and organizing.   It has also been found to regulate, through inhibition and other mechanisms, subcortically generated emotional reflexes.  The PFC and amygdala are linked through bundles of white matter fiber tracts that enable communication between the two areas.  Functional neuroimaging has revealed disease-specific disturbance in each of the three areas of this brain system.

Finding #3:   The functioning of the amygdala is often abnormal in bipolar disorder.  This is especially true of manic states where increased activity is found.  The data is less clear in states of depression and euthymia.  Regarding the latter, some studies report disturbed activity during remission, others show normalized functioning.   Some of these mixed results may be due to the type of task used during the fMRI.  A few studies have shown increased recruitment of the amygdala even during cognitive challenges, such as calculating a sum of numbers.  They imply that bipolar individuals may use the emotional part of their brain, even for more objective considerations.

Finding #4:  Bipolar disorder is also associated with disturbed amygdala size.  This is an example of a structure-function correlation.  While less robust than the functional findings, data suggests that early in the course of illness, the amygdala is actually smaller than found in those without bipolar disorder.  Interestingly, this pattern reverses with age, where bipolar adults are found to have larger amygdalae than their unaffected peers.  We do not yet understand how disturbances in function affect the size of a brain area.

Finding #5:  The prefrontal cortex shows reduced activation in bipolar disorder.  This appears to be true across mood states (i.e., whether a person is depressed, manic or euthymic).  It is most prominent during emotional tasks where the PFC fails to adequately modulate the overactive amygdala.  Think of a surge protector that is not working properly:  Instead of controlling and allowing the smooth flow of electricity, the device malfunctions permitting variable and excessive voltage.

Finding #6:  Neuroprogression.  This term refers to the concept that brain abnormalities irreversibly worsen over time similar to what occurs, for example, in Parkinson’s disease.  This is a question of immense prognostic importance.  To determine this, serial neuroimaging is required, i.e., testing a person in adolescence and again in adulthood, or comparing children at risk, to those in the early vs. late stage of the illness.  There is less data here than with the snapshot studies comparing bipolar adults to those without the disorder at a single point in time.  Nonetheless, the early research suggests that the amygdalae of individuals with bipolar disorder increase in size over time and that the PFC, in contrast, shrinks.  We don’t really know what to make of these observations.  Are they caused by the illness itself?  By the medications used to treat the illness (lithium has actually been shown to cause the amygdala to grow)?  Are the changes irreversible?  Given our uncertainty about this, it may be premature to characterize bipolar disorder as neuroprogressive in nature.

Finding #7:  Connectivity problems.  As mentioned earlier, the most implicated circuit in manic depression includes the amygdala, the PFC, and their connections.  These connections consist of white matter tracts that carry signals between these different brain regions.  fMRI and DTI studies have revealed both functional and structural impairments in these white matter connections in bipolar disorder.  These abnormalities have also been found before the onset of the illness (in children-at-risk, for example) and may thus represent vulnerability markers.

Finding #8:  The functional and structural abnormalities mentioned above are not, by and large, found in schizophrenia.  This supports the idea that bipolar disorder is a distinct illness, with its own neural substrate, and that this brain substrate is specific to the system of circuits that mediate emotional experience.

Conclusion and Cautionary Notes

The increasing clarification of the neural abnormalities associated with bipolar disorder is an exciting and potentially promising development for our field.  Being able to visualize the individual structures and integrated operational circuits that underlie both normal and abnormal mood states is an enormous step forward.    Appreciation of this advance, however, should be tempered by two cautionary notes.  First, describing the functional brain underpinnings of a cognitive or emotional process is not synonymous with defining its cause.  It simply lets us know what regions of the brain are doing at that time.  Discernment of the ultimate cause of an extreme mood state may require additional genetic, cellular, epidemiologic, and psychological investigation.

Second, despite its promise, functional neuroimaging is not yet yielding current clinical application.  In other words, there is no current clinical role for the type of brain scans described above in the evaluation and treatment of someone with bipolar disorder.  For the moment, these tests are used exclusively in research studies.  We expect and look forward to this changing in the very near future.



  1. Blond, B. N., C. A. Fredericks, et al. (2012). “Functional neuroanatomy of bipolar disorder: structure, function, and connectivity in an amygdala–anterior paralimbic neural system.” Bipolar Disorders 14(4): 340-355.
  2. Hafeman, D. M., K. D. Chang, et al. (2012). “Effects of medication on neuroimaging findings in bipolar disorder: an updated review.” Bipolar Disorders 14(4): 375-410.
  3. Schneider, M. R., M. P. DelBello, et al. (2012). “Neuroprogression in bipolar disorder.” Bipolar Disorders 14(4): 356-374.
  4. Strakowski, S. M., C. M. Adler, et al. (2012). “The functional neuroanatomy of bipolar disorder: a consensus model.” Bipolar Disorders 14(4): 313-325
  5. Townsend, J. and L. L. Altshuler (2012). “Emotion processing and regulation in bipolar disorder: a review.” Bipolar Disorders 14(4): 326-339.
  6. Whalley, H. C., M. Papmeyer, et al. (2012). “Review of functional magnetic resonance imaging studies comparing bipolar disorder and schizophrenia.” Bipolar Disorders 14(4): 411-431.