Exploring the mechanisms of action of the antidepressant effect of the ketogenic diet
探討生酮飲食抗憂鬱作用的作用機制

Clinical and neurobiological overlap between depression and epilepsy

Depression and epilepsy are highly comorbid and interrelated, with up to 30% of epilepsy patients being affected by MDD, compared to a much lower (10%) prevalence of MDD in the general population (Błaszczyk and Czuczwar 2016). Other mood disorders, like anxiety and bipolar disorder, have also been investigated in their relationship with epilepsy (Brandt and Mula 2016; Knott et al. 2015). Research has drawn similarities between possible causes of mental disorders and epilepsy at the neural, metabolic, immune, and genetic levels. Both MDD and epilepsy have common cortical structure dysfunction in the hippocampus, neural pathway hyperactivity, and neurotransmitter imbalances involving Gamma Aminobutyric Acid (GABA) and glutamate (Błaszczyk and Czuczwar 2016). Mitochondrial stress and damage can result in both an increased susceptibility to seizures and mood disorder symptomology (Patel 2004; Tobe 2013). Inflammatory markers are overexpressed in both conditions. These interfere with neurotransmitter production and trigger stress pathway overactivity, leading to a brain more susceptible to seizures and depressive symptoms (Mazarati et al. 2017). Finally, initial data suggests a genetic link between temporal lobe epilepsy (TLE) and depression at the Tyrosine Receptor Kinase B (NTRK2) gene (Torres et al. 2017).

Understanding how these diseases co-occur may contribute to more accurate, timely diagnoses and better outcome management (Table 1). Depressive and anxiety symptoms may before, during, and after seizures in patients. Anxiety and stress specifically, is one of the most commonly patient-reported precipitants of seizures (Sperling et al. 2008). They can also be a manifestation of the seizure itself, as in temporal lobe epilepsy, or as a physical or cognitive reaction after the episode. Patients may present as depressed immediately after a seizure due to reactive neural activity inhibition in the frontal lobes (Błaszczyk and Czuczwar 2016). On the other hand, patients can present with affect changes in between seizures and ultimately receive the diagnosis of MDD (Błaszczyk and Czuczwar 2016). Specific tools have been developed to look at mood in those with epilepsy, like the Neurological Disorders Depression Inventory for Epilepsy (NDDI-E) (Gill et al. 2017). Treatment wise, the toxic effects of valproic acid on the mitochondria may even be a potential link between epilepsy patients and MDD development (Silva et al. 2008). Mitochondrial dysfunction has been proposed as a cause of MDD after post-mortem patient investigations (Allen et al. 2018). Although these studies look at genetic causes of mitochondrial dysfunction, valproic-acid induced toxicity could also be considered as a potential etiological link between epilepsy and MDD in patients.
了解這些疾病如何同時發生可能有助於更準確、及時的診斷和更好的結果管理（表 1 ）。患者在癲癇發作之前、期間和之後可能會出現憂鬱和焦慮症狀。具體來說，焦慮和壓力是患者報告的最常見的癲癇發作誘因之一（ Sperling et al. 2008 ）。它們也可能是癲癇發作本身的表現，如顳葉癲癇，或是發作後的身體或認知反應。由於額葉反應性神經活動​​抑制，患者可能在癲癇發作後立即表現出憂鬱（ Błaszczyk 和 Czuczwar 2016 ）。另一方面，患者可能會在癲癇發作之間出現情緒變化，並最終被診斷為 MDD（ Błaszczyk 和 Czuczwar 2016 ）。已經開發出專門的工具來觀察癲癇患者的情緒，例如癲癇神經疾病憂鬱量表 (NDDI-E)（ Gill 等人，2017 年）。在治療方面，丙戊酸對粒線體的毒性作用甚至可能是癲癇患者與 MDD 發展之間的潛在關聯（ Silva 等人，2008 ）。在對患者進行屍檢調查後，粒線體功能障礙被認為是 MDD 的一個原因（ Allen et al. 2018 ）。儘管這些研究著眼於粒線體功能障礙的遺傳原因，但丙戊酸誘導的毒性也可以被認為是患者癲癇和重度憂鬱症之間的潛在病因聯繫。

The characterization of epilepsy, similarly to MDD, has also been migrating from a brain illness to a multi-system condition. For example, both patient populations seem to be at risk for sudden cardiac death (Bauer et al. 2015; Shi et al. 2017). MDD and epilepsy seem to influence the risk, severity, and outcomes of several endocrine, immune, cardiovascular, and metabolic diseases. Studies have found an association between epilepsy and autoimmune conditions like rheumatoid arthritis, cardiovascular events like stroke, and metabolic conditions like hypertriglyceridemia (Nair et al. 2016; Ong et al. 2014; Shmuely et al. 2017; Yuen et al. 2018). MDD has been associated with comorbidities in each of these systems as well (Benros et al. 2013; Ghanei Gheshlagh et al. 2016; Lespérance et al. 2002; Mezuk et al. 2008). The co-presentation of neurological and systemic conditions argues against single-system oriented treatments. If these diseases affect the whole body, so should the treatments.
與MDD類似，癲癇的特徵也已從腦部疾病轉變為多系統疾病。例如，這兩個患者群體似乎都面臨心臟猝死的風險（ Bauer 等人，2015 年； Shi 等人，2017 年）。 MDD 和癲癇似乎會影響多種內分泌、免疫、心血管和代謝疾病的風險、嚴重程度和結果。研究發現癲癇與類風濕性關節炎等自體免疫疾病、中風等心血管事件以及高三酸甘油脂血症等代謝性疾病之間存在關聯（ Nair 等人，2016 年； Ong 等人，2014 年； Shmuely 等人，2017 年； Yuen 等人，2018 年）。 MDD 也與這些系統的合併症有關（ Benros 等人，2013 年； Ghanei Gheshlagh 等人，2016 年； Lespérance 等人，2002 年； Mezuk 等人，2008 年）。神經系統疾病和全身性疾病的共同表現反對單一系統的治療。如果這些疾病影響全身，治療也應該如此。

There seems to be overlap in the efficacy of treatments between disorders as well. Psychotherapies and pharmacologic treatments/programs that specifically target mood are being developed and implemented for epilepsy patients (Błaszczyk and Czuczwar 2016). Neurotransmitter levels are common targets of anti-epileptics and anti-depressants (Kanner 2003). Both selective serotonin reuptake inhibitors (SSRIs), used for depression, and anti-convulsants like valproic acid, used for epilepsy, increase 5-HTT concentrations in the synapse (Baf et al. 1994; Blier and de Montigny 1994). Valproic acid, specifically, has been shown to have mood stabilizing properties (Chiu et al. 2013). Other anti-convulsant drugs target the vagus nerve to stimulate locus coeruleus production of norepinephrine (NE), another common neurotransmitter targeted by serotonin and norepinephrine reuptake inhibitors (SNRIs) (Roosevelt et al. 2006; Sansone and Sansone 2014). Both drug classes have been suggested to affect inflammation, and SSRIs may also lower seizure thresholds (Błaszczyk and Czuczwar 2016; Rapoport and Bosetti 2002; Zou et al. 2018). Overall, common treatment pathways may affect symptomology in both disorders in similar ways.

Depression and epilepsy are highly co-involved at the etiological, diagnostic, treatment, and outcome levels. A deeper understanding of the pathophysiology of one of these illnesses can elucidate new avenues of research for the other. Likewise, integrating treatment options could be useful. The benefits of KD for the management of epilepsy have been demonstrated and replicated (D'Andrea Meira et al. 2019). Understanding how KD improves epilepsy and mood outcomes, and by association, depression, will pave the way for future research and clinical trials.

Effect of KD in mood: evidence from animal and human studies

Mood dysregulation and anhedonia are core characteristic features of MDD. Little is known about the direct effect of KD on mood in MDD and other psychiatric disorders, as few studies have purely evaluated those outcomes. Several animal studies have investigated KD in depression model mice, and the research in humans has generally evaluated mood as a secondary outcome to seizure recurrence. KD has been found to impact each domain of depression: physical, emotional, and cognitive.

Exploring animal models

Animal models have demonstrated the success of KD in the physical and emotional domains. Behaviorally, researchers sought to investigate whether KD could replicate the effects of antidepressants. When given KD, compared to control diet, depressed mice spent less time immobile during the Porsolt test, a learned-helplessness based behavioral assessment where time spent immobile is positively associated with depressive symptomology in mice (Murphy et al. 2004). Mice given anti-depressants exhibit a similar behavioral profile (Murphy et al. 2004). With respect to emotionality, another mouse study found that ketone supplementation alone resulted in decreased anxious behavior during the elevated plus maze, where mice can choose to spend time in either an open or closed arm (Ari et al. 2016). Open arms are meant to stimulate anxiousness, forcing mice with higher anxiety to stay in the closed arms (Ari et al. 2016). Mice supplemented with ketones spent more time in the open arms and less time in the closed arms compared to those on control diets (Ari et al. 2016). KD may have both mood boosting and anxiolytic properties.

KD promotes long-lasting neural changes as well. Intra-uterine exposure alone to KD appears to have an impact on brain development and on both the behavioral and emotional domains of depression (Sussman et al. 2015). Those mice whose mothers were on KD during gestation exhibited less anxiety-like behaviors in the open field test, greater resilience in the forced swim test, and were more physically active than their control diet counterparts. It is important to note that both groups were given standard diet postnatally. These neural changes must have occurred during gestation while the mother was on KD and have sustained throughout the lifespan. A recent study has also looked at non-neuronal changes in rat brains and found that glia undergo morphological changes, especially with respect to branching, in the hippocampi of KD-fed mice compared to control (Gzielo et al. 2019). More needs to be done to investigate how these morphological changes translate into effects on neuronal function, and by extension, behavioral and psychological outcomes. Therefore, while KD may both improve depressive and anxiety-like behaviors in the short term, it could also be restructuring cellular networks, leading to sustained biological changes in the long term.

Clinical

A Cochrane review on epilepsy patients found that KD affected all three domains of psychopathology; patients exhibited less anxiety (emotional), decreased seizure frequency (physical), and greater functioning (cognitive) (Martin-McGill et al. 2018). A recent RCT in epilepsy has concluded that KD can improve mood, functioning, and decrease anxiety independent of seizure control (IJff et al. 2016). It should be noted, though, that the relief from seizure symptomology may be a mediator of the relationship of KD and mood improvement. More research in this area, specifically with patients only suffering from MDD and other mood disorders, will help elucidate this relationship. For example, preliminary case reports in bipolar disorder have reported improved mood in patients lasting up to two years at time of publication (Phelps et al. 2013).

Whether KD can result in the same benefits independent of the presence of a diagnosis of a mental illness is still a subject of debate. Two published studies in healthy individuals have shown contrasting results. Where one study found improvement in mood, another found no significant difference in mood scores between groups (Iacovides et al. 2019). Of note, though, is that McClernon et al. (2007) had participants diet for 24 weeks and found an effect, while Iacovides et al. (2019)'s intervention lasted only three weeks. Whether these timeline differences are of importance is a subject of further investigation. Elucidating how KD exerts its effects would be helpful to guide future research endeavors within each domain of MDD; KD could become part of the portfolio of treatments available to be considered by clinicians and discussed/reviewed with patients (Table 1).

KD and neurotransmitters

Current views on etiology and treatment of mood disorders largely emphasise the role of neurotransmitter regulation (Delgado 2000; Hirschfeld 2000). Conceptualizing depression simply as a decrease in monoamine levels could certainly be an oversimplification of this condition. Current pharmacologic treatments like SSRIs are based on the monoamine hypothesis and aim to increase serotonin levels - for example, by inhibiting its reuptake from the synapse (Dupuy et al. 2011). However, only 30% of patients respond to first line treatments, thus simply raising neurotransmitter concentrations in the synapse may not be sufficient to treat MDD (Sinyor et al. 2010). Recent theories have expanded such perspectives and highlight the fact that the same neurotransmitter may have different effects, depending on the receptor it binds (Werner and Coveñas 2010). For example, serotonin elicits an inhibitory effect when it binds 5HT1A receptors, and an excitatory effect upon 5HT2A binding (Carhart-Harris and Nutt 2017).

In their scoping review, Werner and Coveñas (2010) highlight that a combination of underactivity or overactivity in separate cognitive networks may be a better way of understanding the neural circuitry of depression. The default mode network (DMN), or brain areas active at rest, seems to be overactive and overconnected in the brain of those with MDD compared to healthy individuals (Raichle 2015; Whitfield-Gabrieli and Ford 2012). These interconnections may be at the root of the ruminative and moody phenotypes of a patient with MDD, especially since the DMN seems to be more connected with the affective mode network and limbic emotional processing areas in these patients (Palmer et al. 2015; Sheline et al. 2010). KD may be the intervention we need to alter the signaling along these different networks because of its broad effects across several neurotransmitter systems.

Many monoamine and amino acid neurotransmitters appear to be affected and play a role in the effect of KD on the brain. Rat studies have found that KD's effectiveness in epilepsy may be mediated by norepinephrine (Szot et al. 2001). NE inhibition is a common anti-convulsant mechanism. Rats on KD had seizure latencies of similar length to rats whose NE synthesis had been knocked out (Szot et al. 2001). Both groups had seizure latencies that were significantly longer than normal rats on regular chow (Szot et al. 2001). Since KD and NE KO mice had a similar behavioral phenotype, this suggests that KD has an inhibitory effect on norepinephrine transmission, effectively mimicking a common mechanism of action of anti-convulsants. SNRIs also target the norepinephrine system. This could elucidate another common pathway between the effectiveness of KD for epilepsy and its potential benefits in depression. One human study did not find norepinephrine metabolites change significantly in the KD group, emphasizing the need for further research (Dahlin et al. 2012).

Other human studies have suggested that dopamine, serotonin, and glutamate levels decrease in KD groups, while GABA levels increase (Dahlin et al., 2005, 2012; Juge et al. 2010). Dopamine, serotonin, and glutamate all have excitatory effects in the brain (Amato 2015; Ko and Strafella 2012). While typical antidepressants aim to raise monoamine levels, not all patients respond to this treatment (Khan and Brown 2015). By decreasing the levels of these excitatory neurotransmitters, KD may be an effective way of targeting the overactive affective networks elucidated in recent research. Research previously discussed in this review has highlighted the beneficial effects of KD on mood. Decreasing the amount of pathological neural hyperactivity between the DMN and limbic systems may be at the core of KD's mechanism of action.

The increase in GABA levels supports this hypothesis. Returning to the commonalities between depression and epilepsy, seizures are often characterized as the result of neural over-activity (Bromfield et al. 2006). Anti-convulsants target GABA to dampen this activity and reduce seizure incidence (Czapiński et al. 2005). Epileptic patients on KD have both a decrease in seizure activity and a large increase in GABA metabolites (Dahlin et al. 2005). This presents another way by which KD could return DMN activity among patients with MDD to normal levels. In the long term, through neuroplasticity, this could potentially target the over-connectivity also seen in the brains of those with MDD and lead to long-term change (Andrade and Rao 2010). By potentially targeting hyperactivity along maladaptive neural networks, KD may be working on the broader scale necessary for individuals whose depressive symptoms do not respond to current pharmacological options.
GABA 水準的增加支持了這個假設。回到憂鬱症和癲癇之間的共同點，癲癇發作通常被認為是神經過度活動的結果（ Bromfield 等人，2006 ）。抗驚厥藥物以 GABA 為標靶來抑制這種活性並減少癲癇發作（ Czapiński 等，2005 ）。服用生酮飲食的癲癇患者癲癇發作活動減少，但 GABA 代謝物大量增加（ Dahlin 等，2005 ）。這提供了 KD 可以使 MDD 患者的 DMN 活性恢復到正常水平的另一種方法。從長遠來看，透過神經可塑性，這可能會針對重度憂鬱症患者大腦中出現的過度連接，並導致長期變化（ Andrade 和 Rao 2010 ）。透過潛在地針對適應不良神經網路的過度活躍，KD 可能會在更廣泛的範圍內發揮作用，以治療那些憂鬱症狀對當前藥物選擇沒有反應的個體。

KD and cell energetics KD 和細胞能量學

Metabolic anomalies that are expressed in mood pathologies may constitute a potential link between cell energetics and MDD (Allen et al. 2018). In humans, Fattal et al. (2006) found 19 different case reports of patients with mitochondrial diseases co-presenting with psychiatric illness. Mitochondrial disorder modeling in mice seems to induce behavior indicative of mood dysfunction (Kasahara et al. 2006). In another animal model, in which depression was induced using a chronic stress paradigm, rats developed a reduced craving for sweet food (the rodent form of anhedonia) (Rezin et al. 2008). The researchers found that the depressed rat brains had complex I, III, and IV of the electron transport chain inhibited in the cortex and cerebellum (Rezin et al. 2008). These studies propose a bidirectional relationship between mitochondrial function and mood. KD could capitalize on this relationship through mitochondrial-mediated improvements in mood.

Targeting the mitochondrial can have an impact on mood. The use of N-acetylcysteine (NAC) in mood disorder patients is especially promising (Berk et al. 2019). NAC has been shown to be neuroprotective while targeting mitochondrial-based energy defects (Banaclocha 2001). In aging mice, NAC administration resulted in increased complex IV and oxidative phosphorylation activity (Martínez Banaclocha 2000). Its antioxidant properties reduce reactive oxygen species (ROS) production and inhibit subsequent damage to the cell and mitochondria (Halasi et al. 2013). When NAC was given to bipolar and patients with MDD over 16 weeks, significant improvements in symptoms were found (Berk et al. 2019). This illustrates the potential benefits of KD in treating psychiatric illness at the metabolic level.

Studies in mice have found that KD directly affects cellular energetics. After following KD for 33–58 h, the brain undergoes a metabolic ‘switch’ from glucose to ketone metabolites as its main energy source (Harvey et al. 2018). BHB, the major ketone metabolite, has been attributed as the main player in the resulting cellular changes (Maalouf et al. 2009). BHB targets the mitochondria and renders it more resilient to stress by increasing glutathione levels, a known antioxidant, reducing its production of ROS, and increasing UCP expression (Jarrett et al. 2008). During ATP production, protons leak into the mitochondrial matrix through the inner membrane. UCPs interfere with this process and decrease ROS production. Therefore, similar to NAC, ketone bodies reduce ROS production by increasing UCP levels (Sullivan et al. 2004). This all leads to a net increase in mitochondrial biogenesis, greater cellular respiration, and higher ATP production rates (Hasan-Olive et al. 2019; Sullivan et al. 2004). KD's neuroprotective properties decrease ROS production, raise antioxidant levels, and stimulate mitochondrial biogenesis (Bough and Rho 2007; Hasan-Olive et al. 2019; Sullivan et al. 2004). Not only do mood and metabolic disorders often co-present but improving mitochondrial health may have an impact on mood as well. KD's neuroprotective properties could help reduce cellular stress on the mitochondria, and in doing so, alleviate symptoms of MDD.

KD and inflammation

Although still relatively in its early stages of investigation, current research has found inflammatory markers to be abnormally high in depressed brains (Amodeo et al. 2017; Barnes et al. 2017; Zou et al. 2018). In fact, a bidirectional association has been found between inflammatory disorders and depression, where a patient is at increased risk for one by having the other (Amodeo et al. 2017). A similar relationship has been found in their respective treatments. Anti-inflammatories like NSAIDs seem to improve depressive symptoms in patients with MDD, while antidepressants have been shown to reduce inflammatory marker levels in the same patient population (Anglin et al. 2015; Hannestad et al. 2011). Therefore, if KD reduces inflammatory marker levels, this could potentially be another source for symptom improvement in patients with MDD.

Several animal studies have investigated KD's anti-inflammatory properties. Beginning from outside the brain, after being on KD for four weeks, rats that are exposed to painful stimuli seem to express less inflammatory blood markers than their normal chow counterparts (Ruskin et al. 2009). KD has been shown to decrease reactive inflammatory marker expression in the brain after an external insult (Dupuis et al. 2015). Other researchers have simulated inflammatory disease states in animal brains, be it in MPTP mice models of Parkinson's disease (PD) or encephalomyelitis mice models, and reported that KD significantly reduced inflammatory markers and improved disease phenotype (Kim et al. 2012; Yang and Cheng 2010). It is suggested that the BHB metabolite of KD may be the mediator of the neuroprotective effects of KD by lowering IL1B, IL6, TNF-α, and other chemo and cytokine levels (Yang and Cheng 2010). These are the same markers involved in depression that KD may be targeting, therefore leading to the improvements in mood symptoms (Amodeo et al. 2017).

Inflammation has also been attributed as a potential source of epilepsy and seizure propagation (Mazarati et al. 2017; Vezzani et al. 2011). Thus, the mediation of inflammatory pathways could be another path by which KD potentially improves one disease state and subsequently impacts the other. In an epilepsy mouse model, KD inhibited pro-inflammatory marker increases during kainic acid (KA) induced seizure trials when compared to control diet mice (Jeong et al. 2011). The authors reported markedly diminished levels in TNF-α, NF-κB, and COX2 levels in the KD mice, each of which has been implicated in excitotoxicity and neuronal cell death in the context of epilepsy (Jeong et al. 2011).

The extent to which KD modulates inflammatory signaling in depression is still an open question - both in humans and in animal models. The immune system seems to have a role in both the etiology and treatment of depression. Specifically, increases in cyto- and chemokine levels could be a potential mediator of the pathology seen in depressed brains. Research has shown that KD can dampen pathological inflammatory responses in epilepsy and neurodegenerative diseases in animal models. The investigation of the effects of KD on inflammatory markers in depression is long overdue – particularly their contribution to the improvements in depressive traits observed so far with the adoption of KD.

KD and the microbiome

The link between the brain and the GM that inhabit the digestive system is growing stronger with recent evidence in psychiatric and neurological research (Thomas et al. 2017). Human and animal studies have found differences in the abundance and diversity of GM in those with clinical disorders, including MDD (Foster and McVey Neufeld 2013). The families of bacteria that are more predominant over others in the gut can impact mood and inflammation in the body (Dinan and Cryan 2017; Foster and McVey Neufeld 2013). A patient's diet, medications, and disease status can also have a reciprocal impact on the GM (Cresci and Bawden 2015). This bidirectional relationship is important when considering KD and its impact on MDD, potentially through GM changes. Human and animal studies show that KD alters the microbiome by: 1) reducing diversity, 2) reducing abundance, and 3) altering the composition (Newell et al. 2016).

Moods induced GM changes, and vice versa. When mice are exposed to chronic stress, their GM changes in comparison to healthy controls (Bailey et al. 2011). These mice exhibit a larger abundance of Clostridium and less Bacteroides than their non-stressed counterparts (Bailey et al. 2011). Similarly, in humans, GM analysis has found a significantly decreased representation of the Bacteroidales class in individuals with MDD compared to healthy controls. These studies suggest that both external and internal conditions can alter GM, leading to the hypothesis that successful treatments that remedy these conditions may also have a positive impact on GM.

The anti-microbial properties of SSRIs have an impact on GM (Lukić et al. 2019; Macedo et al. 2017). Researchers suggest that changing the dominant bacterial profiles in the gut modulates inflammatory pathways, and in turn supports the hypothesis that inflammation may play an important role in mental health. Jiang et al. (2015) compared the GM between healthy controls, MDD patients who responded to treatment, and MDD patients who continued to have active depressive symptoms. The actively depressed group saw a greater representation of Enterobacteriaceae and Fusobacterium and less Bacteroidaceae in their GM when compared to controls. Conversely, those with MDD that were responsive to treatment exhibited higher levels of Bacteriodetes and Proteobacteria compared to controls. Both MDD groups had less Lachnospiraceae and Ruminococcaceae representation in their GM than control. The authors conclude that certain bacterial profiles, like higher Bacteriodetes levels, may serve as a useful biomarker for patients that will respond well to treatment.

These results are supported in the epilepsy literature that focused on the effects of KD on GM and treatment response. The link between GM and epilepsy can be seen when chronic stress paradigm-induced GM changes affects mouse model susceptibility to seizure kindling and propagation (Medel-Matus et al. 2018). These effects carry forward when transplanting the GM from stress to non-stress mice (Medel-Matus et al. 2018). The latest research in mice has found that KD can reduce seizure thresholds compared to control diets (Olson et al. 2018). Further, transplanting KD GM into germ free (GF) mice transfers the resilience to seizure activity as well. This is promising research that suggests the GM to be a crucial component of KD's mechanism of action in epilepsy, and perhaps in other brain disorders as well.

Several studies have investigated the effect of KD on the GM in epilepsy patients taking anti-convulsants. Healthy controls consistently exhibit higher levels of the Bacteroides genus compared to epilepsy patients (Spinelli and Blackford 2018; Xie et al. 2017; Zhang et al. 2018). Cronobacter, Clostridiales, Clostridium, and Ruminococcaceae populations are higher in patients compared to controls before KD (Spinelli and Blackford 2018; Xie et al. 2017; Zhang et al. 2018). Patients are classified as either treatment responders or non-responders after treatment with KD. Patients that respond to KD develop GM profiles that are more similar to healthy controls by decreasing their expression of the bacteria previously mentioned and increasing their representation of Bacterioides. Meanwhile, KD non-responders saw an increase in their expression of Clostridiales, Ruminococcaceae, and Lachnospiraceae.

Across the epilepsy and mood research, it seems that greater expression of Bacterioides and Bacteroidetes in the GM leads to better mental health outcomes. Further, KD appears to specifically increase the levels of these bacteria in those patients that respond to treatment. Researchers suggest that Bacteroides levels could be a useful biomarker to monitor treatment response. If MDD patients that respond to treatment typically have higher levels of Bacteroides, and KD increases Bacteroides levels in patients, future research should investigate whether a direct link exists between KD, GM, and depression outcomes in patients. KD is a metabolic intervention whose impact on GM levels and bacterial representation could have a positive impact on those with treatment-refractory MDD (See Figures 1 and 2).

Figure 1:

Illustrating the current understanding of the potential effects of KD on MDD. RDoC, or Research Domain Criteria, is a framework created by the NIH that characterizes the features of mental illnesses and helps guide research efforts in the field (“NIMH 2008 » Research Domain Criteria (RDoC),” n.d.).

Figure 2:

The changing focus of psychiatric treatments over time.