In BD, patients suffer from severe mood fluctuations which range from major depressive episodes (MDE) to manic or hypomanic exacerbations. Recognized by Kraepelin and described in the 8^th edition of his textbook, mixed states have a high prevalence rate in this ailment.1 Since that era, the definition and conceptualization of mixed states has varied, but with the introduction of the Diagnostic and Statistical Manual's 5^th edition (DSM-5) in 2013, this notion has crystallized and broadened.2 In the preceding DSM-IV-TR, mixed episode was defined in the context of BD type I as an occurrence typified by the concurrent presence of full manic and depressive phenomena in an individual for at least seven days. This narrow definition was radically modified in DSM-5, with its redefinition that the presence of only three symptoms of opposite mood polarity was sufficient to qualify for the specifier “with mixed features”. Moreover, this specifier could be used for a manic episode in BD type I, hypomanic episode in BD type I or II, as well as for an MDE in major depressive disorder (MDD).3 The paradigm shift reflected the accumulated research evidence which emphasized the importance of mixed symptoms in the clinical, functional, and therapeutic spheres of affective disorders.4 Fig. 1 gives an illustrative account of the current understanding of mixed states throughout the entire spectrum of mood disorders and is reflective of the widening of this concept. Table 1 provides an overview and diagnostic features of incumbent affective states according to DSM-5.

Within the framework of DSM-5 the extension of the perception of mixed states implies that in an affective episode, clinicians are able to diagnose patients with only few concomitant symptoms of opposite mood polarity. Essentially, this shows that bipolar and MDD patients are more likely to be considered as having mixed episodes, and while this is the modern rendition of a century old idea espoused by Kraepelin, it has significant ramifications for the pathogenesis and treatment of these ailments. Approximately 40% of patients have mixed episodes, but this figure may be much higher within the context of DSM-5.5 For the improved management of these conditions, it is therefore necessary to cultivate a better understanding of the neurobiology and longitudinal course of mixed states.

In the perspective of bipolar disorder, recurrent euphoric mania occurs in a small subset of cases comprising about 30% of the patients, whereas dysphoric or mixed mania is much more prevalent.6 Bipolar patients tend to have repeated episodes of the same type, so that those suffering from mixed episodes are prone to exhibiting the same type of symptoms time and again. They suffer from greater comorbid conditions, chief among which are anxiety spectrum disorders and ailments due to substance misuse. Furthermore, due to the presence of depressive phenomena in activated states, these patients are also at greater risk of completed suicide.7 Another factor negatively influencing the prognosis is susceptibility to rapid cycling, in which patients have brief if any euthymic periods and a persistent symptomatic state is evident throughout the course of the illness. These facts allude to the pernicious nature of mixed symptoms, so that the treating physician must be alert to their presence and be prepared to institute therapeutic strategies which are effective, well tolerated, and have a good safety profile.8 Fig. 2 provides a pictorial outline of the burden imposed by mixed affective symptoms on the illness course in bipolar disorder, with a depiction of the long term adverse consequences.

While the onset of BD is frequently in adolescence or the early teenage years, the first manifestations are varied and often confusing. A young adult can present with an initial episode of melancholic depression, or alternatively with a disturbance of conduct like oppositional defiant disorder or attention deficit hyperactivity disorder. There may be psychosocial perturbation with emotional dysregulation, substance misuse and involvement in risky behaviors. In this regard, family history may be suggestive, with the presence of bipolar disorder in 1^st degree relatives, but many symptomatic youth have close family members suffering from such diverse neuropsychiatric conditions as anxiety spectrum disorders, substance dependence, or psychotic spectrum disorders.9 With the present state of knowledge there are still no validated biomarkers for BD, so that the treating physician has to take a longitudinal perspective and manage such cases empirically. The course of BD is malignant in a substantial minority of cases, and as pharmacotherapuetic options are limited and there is a presence of treatment refractoriness, it is a serious issue.10 Repeated affective episodes of increasing severity have a pervasive, negative effect on all aspects of functioning, with adverse sequelae in the biopsychosocial realm. In this scenario, it is essential to inculcate better knowledge of the neurobiological underpinnings of affective exacerbations, particularly the pathogenesis of manic and depressive episodes, the homeostatic disturbances characterizing mixed states, and neurochemical alterations driving the switch process.

Mood disorders are exemplified by disruption in the sleep-wake schedule, disturbance in daily activity levels, and irregularity of meal times, signifying a profound disorder in the circadian system.11 While these aspects of affective illnesses have recognized since the mid-20^th century, research in the past few decades has progressively clarified the etiopathological mechanisms underpinning this correlation.12 It has now become obvious that circadian rhythm disturbance is at the core of mood disorders and this section of the review will outline the findings from the extant literature in this regard, with particular reference towards the induction and perpetuation of mixed states.

It has been demonstrated that in patients with BD, mood exacerbations are provoked by alterations in the intensity of light and also follow seasonal patterns.13 Further, bipolar subjects show irregularities in key biological processes that are under circadian control, for example sleep, diurnal activity, hormone levels and body temperature.14 Mood stabilizing medications such as lithium act to reinstate some of these disrupted rhythms by producing a robust phase-delay in daily oscillations and increasing rhythmic amplitude, which may be the key to their therapeutic effects.15 With the identification and replication of the individual genes that constitute the molecular clock, it is now possible to investigate the molecular mechanisms that underlie the association between the circadian system and mood disorders. Additionally, new insights are developed that inform us about the pathogenesis of different phenotypes within affective disorders, such as mixed states and their relationship to the dysregulation of the circadian machinery.16 A brief description of the molecular clock follows to clarify the sub-cellular events that epitomize the daily, rhythmic oscillations across the mammalian species.

The main molecular clock is housed in the supraschiasmatic nucleus (SCN) in the hypothalamus, and is formed of a transcription-translation circuit which oscillates over the duration of about 24 hours, without the contribution of external environmental cues. The chief activator of transcription is comprised of a dimerized complex of Circadian Locomotor Output Cycles Kaput (CLOCK) protein and Brain and Muscle ARNT-like Protein 1 (BMAL1). This molecular composite binds to Enhancer-box sequences in the promoters of many genes including the Period (Per) and Cryptochrome (Cry) genes. The PER and CRY proteins are translated in the cytoplasm and are phosphorylated by the priming enzyme, casein kinase 1Є before being targeted by glycogen synthase kinase 3β, causing alterations in their viability, relationship, and capacity to enter the nucleus. Upon DNA binding, these inhibit the actions of the CLOCK/BMAL1, thus forming a negative feedback loop. Furthermore, there is another closely related loop in which the CLOCK/BMAL1 activate the transcription of orphan nuclear receptor genes Rev-erbα and Rorα which act to repress and activate the transcription of Bmal1 and Clock, respectively through their interaction with the ROR elements. As alluded to above, many controlling kinases, phosphatases, and secondary feedback loops act on the molecular clock, contributing to further intricacy in the circadian apparatus.17 Significantly, circadian transcription factors are involved in the regulation and functioning of several other clock-controlled genes, which partake in a whole range of homeostatic actions in every body system. In essence, it is now believed that approximately 50% of the mammalian genes are expressed rhythmically, which participate in a wide array of biological and behavioral functions.18

In BD, every patient shows signs of manic and depressive symptoms to a greater or lesser extent and varying pathophysiological mechanisms expectedly trigger these affective changes. There is a whole gamut of presentations, as some patients have rapid fluctuations in symptomatology extending over days and weeks, while other more severe cases have concomitant manic and depressive manifestations that alternate over a matter of hours, a phenotype sometimes referred to as ultra-rapid cycling bipolar disorder.19 For effective treatments, the mechanisms underlying switching between states need to be discovered, and this aspect is aptly regarded as the “holy grail of BD research”. In this respect, important insights have been inculcated by studying the switch process brought about by pharmacotherapuetic and chronobiological measures employed for the management of mood disorders.20

Patients with bipolar depression treated with different classes of antidepressants exhibit varying frequency of change over into manic or hypomanic states, and mixed episodes are also reported to occur with higher frequency in such cases. Long term treatment with tricyclic antidepressants (TCAs) has time and again been associated with producing this affective change over in up to 70% of recipients.21 TCAs mainly act to block the re-uptake of serotonin and norepinephrine from the synaptic cleft and these combined blockers produce manic/hypomanic switches with greater regularity than selective serotonin re-uptake inhibitors.22 Bupropion, a dual norepinephrine/dopamine re-uptake blocker also causes switches in about 20% of treated cases, and this occurs in spite of co-treatment with mood stabilizers. It therefore appears that increasing synaptic levels of norepinephrine or dopamine via re-uptake inhibition likely provokes change over from depression to hypomania or mania, as well as results in the induction of mixed states even when mood stabilizers are employed adjunctively.23

Alternatively, directly enhancing norepinephrine/dopamine levels also results in mania-like behavior. For instance, therapeutic administration of L-dopa to Parkinson's disease patients can provoke mania like behaviors such as gambling and increased risky activities.24 Furthermore, indirect dopamine/norepinephrine augmentation by amphetamine, a combined re-uptake inhibitor, can cause manic/hypomanic exacerbations in euthymic bipolar patients, and instigate behavior resembling mania in well individuals. Confirmation of these effects is derived from research which shows that alterations of dopamine/norepinephrine re-uptake sites can modify the behavioral outcomes of amphetamines.25

Tyrosine hydroxylase (TH) is the rate limiting enzyme in dopamine biosynthesis so that administration of the TH inhibitor α-methyl-para-tyrosine (AMPT) causes depletion of catecholamines, and reduces symptoms of mania while increasing depressive manifestations. In one study, hypomania was induced in euthymic bipolar patients after revival from a catecholamine reduction caused by AMPT, making it apparent that the regulation of catecholamines with over-activity is important for switching to an activated state, whereas the opposite may be true for depression.2627

There is much less information about pharmaceutical agents that can cause switching to a depressive episode. Physostigmine, a medication employed to enhance memory in Alzheimer's disease is an acetylcholinestrase inhibitor which prevents the breakdown of acetylcholine, thus ultimately increasing its synaptic levels. It can decrease manic symptoms in BD cases, moreover in healthy individuals and remitted cases this medication can produce a state of depression. To extend these findings further, recent work has shown that enhanced acetylcholine levels are seen in subjects with both unipolar and bipolar depression.28 Additionally, there is change in the expression of cholinergic receptors in bipolar subjects, giving credence to the notion that elevated acetylcholine levels purportedly induce a depressed state in BD patients.29

The working of other neurotransmitters has been studied to investigate the phenomenon of switching in BD. Both lamotrigine and riluzole are inhibitors of glutamate release and neither of these has been associated with the switch process. Particularly, the former medication is frequently used in the depressed and maintenance phases of BD, but studies have repeatedly shown that lamotrigine is not associated with switching into manic/hypomanic states, does not induce mixed episodes, nor does it lead to rapid cycling.30 Intravenous administration of ketamine, an NMDA receptor antagonist is an emerging treatment for depression, including bipolar depression. No incidence of ketamine induced switching in depressed bipolar patients has been reported, making it unlikely that glutamatergic signaling is involved in the switch process.31

Fig. 3 provides an illustrated adaptation of the neurochemical imbalance hypothesis of bipolar disorder and emphasizes the continuum in symptomatology during the course of the illness.

In euthymic, unmedicated patients, the switch into mania has been documented with irregular sleep patterns, along with increase in catecholamine release.32 Complete lack of sleep can result in manic exacerbation in up to 30% of bipolar patients, and can also cause mania like behavior in healthy controls.33 These observations support the social rhythm disruption or social zeitgeber theory of mood disorders, which speculates that a disturbance in social/biological rhythms can induce affective episodes in susceptible people. Such disruptions can include stressful life events, irregular daily schedules and chaotic life style with disarray in sleep and meal times, while circadian disturbances caused by shift work, transmeridian travel, etc can also precipitate mood episodes.34 Hence, circadian rhythm irregularities are a major contributor in the pathogenesis of BD, whereas all effective treatments essentially resynchronize the biological clock.

In this vein, it must be understood that changes in seasons are among the most important environmental triggers for precipitation of mood episodes.35 In seasonal affective disorder (SAD), vulnerable individuals experience depressive exacerbations during autumn and winter months, corresponding to shortening of the day. Early morning bright light therapy has a phase advancing effect and is therapeutic in SAD, however it is also known to cause switching into mania, as well as inducing mixed states.36 Further, bipolar patients experience manic episodes at a higher rate during spring and summer months with the lengthening of the period of the day.37 Significantly however, seasonal variations in BD are correlated to a positive family history, and thus the illness has a heritable component making this a genetic-based condition with an affirmative history of the disorder in family members in 60-80% of cases.

Several genome wide association studies (GWAS) have been undertaken in BD but statistically significant correlations have not been forthcoming. This may be because BD is a diverse condition and the patient populations studied were too heterogeneous to provide meaningful results. As such, studying sub-sets of patients with phenotypic similarity can be more informative. In a GWAS on subjects self-proclaiming as having seasonal recurrences of mania versus non-seasonal bipolar cases, an SNP rs41350144 lying within the intron of neurofibromin 1A gene on 1p31 was found to have a statistically significant association in the former sub-population.38 More significantly however, genes involved in the circadian mechanism have also been implicated in BD. In one study, a causal relationship concerning SNPs in 10 CLOCK related genes which included ARNTL, CLOCK, CRY2, CK1Є, DBP, GSK3β, NPAS2, PER1, PER2 and PER3 were investigated. Haplotype analyses in ARTNL and PER3 genes were found to be meaningfully connected to BD in the 159 families included in the study.39 There are several other studies that have identified polymorphisms in the circadian apparatus using candidate gene approaches, whose detailed description is beyond the scope of this review. However, it can be argued that circadian rhythm abnormalities are the cause rather than the effect of mood disturbances.40

Importantly, rodent models with genetically engineered mutants have shed light on the involvement of circadian genes in mood disorders. Mice created by inducing an alteration in the Clock gene (ClockΔ19) have a phenotype comparable to mania in BD. Intriguingly, when these animals are administered the prototype mood stabilizer, lithium, their behaviors in large part normalize resembling those of wild type mice.41 Mutations in other CLOCK genes result in similar behavioral phenotypes, a prime example being of transgenic mice overexpressing GSK3β exhibiting manic like profiles.42 Utilizing ClockΔ19 mice, revealing experiments have shown the link between circadian genes, neurotransmitter functional control, and behavioral phenotypes of high clinical relevance.

In an attention-grabbing study mutant mice were engineered by RNA interference using short hairpin RNA (shRNA) and viral utilized gene transmission to silence Clock expression exclusively in the ventral tegmental area (VTA). The animals were next subjected to a range of behavioral, molecular, and physiological measures. It was discovered that disabling Clock specifically in the VTA resulted in hyperactivity and reduction in anxiety-related behaviors akin to the profile of the ClockΔ19 animals. Interestingly, VTA exclusive Clock silencing also caused a significant rise in depression-like behavior, generating a largely manic-mixed picture. Furthermore, VTA knock-down of Clock also changed circadian period and amplitude, implicating Clock in the VTA in the control of daily oscillations. VTA dopaminergic neurons carrying the Clock shRNA fired at a greater rate than the non-transfected neurons. Clock disabling altered the expression of several ion conduits and dopamine-associated genes in the VTA which could be responsible for the manifest phenotype in the modified strains. As a whole, these observations implied a major function for CLOCK in the VTA in the control of dopamine neurotransmission, expression of behavioral profile reminiscent of mixed states, and dysregulation of circadian rhythms.43

The above mentioned findings were further extended in a more recently published study which again utilized the ClockΔ19 mice having a polymorphism in the CLOCK gene. The latter is a core regulatory constituent of the circadian transcription machinery and Clock mutant mice display increased activity, more reward seeking and reduced anxiety and depression like behaviors, resembling mania in humans. Mice are nocturnal animals and genetically modified strains express this phenotype during daytime; their period of inactivity. Concomitant with this profile, there is increased firing of VTA dopaminergic neurons, which is in line with altered mesolimbic dopaminergic activity observed in human mania. In an elegantly designed study, the interconnection between circadian gene interference and rapid mood cycling/switching was demonstrated. ClockΔ19 mice were used as these exhibited rapid changes in behavior with an intense, manic-like profile seen throughout the day subsequent to a phase of calm at night. Mood oscillations corresponded to atypical daytime rises in VTA dopaminergic neuronal firing, tyrosine hydroxylase (TH) activity and dopamine biosynthesis. The investigators created a novel optogenetic stimulation prototype in wild type (WT) mice, in which the genetically unaltered strains were photo-stimulated (ventral tegmental area only) in a chronic fashion for 7 consecutive days (1 hr/d) followed by behavioral and physiological measurements. The WT mice exhibited same rapid cycling behavioral profile as ClockΔ19 strains and this phenotype persevered for up to 2 weeks after termination of optic stimulation. Further, time-dependent inhibition of TH activity by alpha-methyl-para-tyrosine during the day reversed manic-like behaviors in both types of experimental animals. Moreover, it was demonstrated that the CLOCK protein was a negative regulator of TH gene transcription, disclosing an original molecular mechanistic link between the circadian apparatus and mood switching, and shedding light on the pathogenesis of mood alterations in BD.44

The emerging evidence with respect to the pathogenesis of mood switching and induction of mixed symptoms in BD is schematically depicted in Fig. 4.

There are an increasing number of studies on the application of different cerebral stimulation techniques in the treatment of mood disorders. These include ECT, repetitive transcranial magnetic stimulation, vagus nerve stimulation, and deep brain stimulation. With regards to effectiveness, the existing literature is most supportive of ECT which emerges as a rapidly effective measure in intractable cases who are not showing response to manipulation with existing psychopharmacological agents. ECT is a valid treatment option in actively suicidal patients, catatonic cases, subjects with refractory mixed states and those exhibiting persistent, severe affective symptomatology poorly responsive to pharmacotherapy.72 Table 2 provides a summary of effectiveness of treatment measures on mixed mood states in bipolar disorder.