Evolution in the Treatment of Psychiatric Disorders: From Psychosurgery to Psychopharmacology to Neuromodulation

The treatment of psychiatric patients presents significant challenges to the clinical community, and a multidisciplinary approach to diagnosis and management is essential to facilitate optimal care. In particular, the neurosurgical treatment of psychiatric disorders, or “psychosurgery,” has held fascination throughout human history as a potential method of influencing behavior and consciousness. Early evidence of such procedures can be traced to prehistory, and interest flourished in the nineteenth and early twentieth century with greater insight into cerebral functional and anatomic localization. However, any discussion of psychosurgery invariably invokes controversy, as the widespread and indiscriminate use of the transorbital lobotomy in the mid-twentieth century resulted in profound ethical ramifications that persist to this day. The concurrent development of effective psychopharmacological treatments virtually eliminated the need and desire for psychosurgical procedures, and accordingly the research and practice of psychosurgery was dormant, but not forgotten. There has been a recent resurgence of interest for non-ablative therapies, due in part to modern advances in functional and structural neuroimaging and neuromodulation technology. In particular, deep brain stimulation is a promising treatment paradigm with the potential to modulate abnormal pathways and networks implicated in psychiatric disease states. Although there is enthusiasm regarding these recent advancements, it is important to reflect on the scientific, social, and ethical considerations of this controversial field.


INTRODUCTION
The management of psychiatric disorders is challenging and often requires a multimodal approach to diagnosis and treatment. There is a rich history of innovation in the field, driven by scientists, physicians and surgeons. In particular, the neurosurgical treatment of psychiatric disorders has a long and tumultuous history fraught with controversy. However, the legacy of "psychosurgery" has also given rise to the development of modern standards for research and ethics, and has fostered a deeper understanding of the pathophysiology of human behavior. Although there is evidence of psychosurgery spanning multiple thousands of years throughout human history, the most exciting, but also provocative developments have been within the past century, attributable to the combined efforts of scientists and physicians. In the 1950s, ablative surgery fell out of favor due to the rise of effective pharmacology and intense professional and public criticism, although research and practice continued with more rigorous standards. Although the modern treatment of psychiatric disorders is primarily medical, the high incidence of treatment resistance and failure has fostered a renewed interest in surgical treatments with a non-ablative focus.
The aim of this article is to provide a brief chronological overview of the treatment of psychiatric diseases from ablation to pharmacology to neuromodulation. Despite a history of controversy, interest in the potential of surgery for psychiatric disorders has endured and even increased within the past few decades, primarily driven by the proliferation and success of neuromodulation and by improvements in structural and functional neuroimaging. It is important to view developments in psychosurgery in the context of the historical and current understanding of the neurobiology and pathophysiology of consciousness and behavior, the available treatments for psychiatric disorders, and the adherence to (or lack thereof) research ethics.

A HISTORICAL PERSPECTIVE The Origins of Psychosurgery
The earliest evidence of presumed psychosurgery has its origins in the Neolithic era; several skulls from this period have been identified with areas of trephination and evidence of healing (Stone and Miles, 1990;Alt et al., 1997), suggesting that these early procedures were likely performed with therapeutic intent. Although concurrent fractures have been identified in some specimens, numerous other skulls bear no obvious signs of trauma (Stone and Miles, 1990). It has been hypothesized that early trephination was performed for ritualistic or spiritual purposes, with the intent to treat manifestations of headaches, epilepsy, and mental illness (Robison et al., 2013). During the classical era, detailed guidelines regarding trephination, primarily for trauma, were outlined in the Hippocratic text "peri ton en cephali traumaton, " translated to "On Head Wounds" or "On Injuries of the Head" (Dimopoulos et al., 2008). Psychosurgery was also depicted in the medical literature and artwork of the Renaissance era, notably in "The cure of folly or the operation for the stone, " a painting by Hieronymus Bosch referring to the belief that madness was caused by a physical stone within the brain (Salcman, 2006).
The practice of psychosurgery was then largely absent from Western medicine for a few hundred years until its resurgence in the modern era. Throughout the 1800s, new insights into functional neuroanatomy and neurophysiology laid the foundation for renewed interest. In 1819, Franz Joseph Gall published his treatise on phrenology, which suggested that the brain possessed discrete functional regions (Simpson, 2005).
Although phrenology was flawed and ultimately discredited, the idea of neurological functions having an anatomical correlate was expanded on by seminal work on the localization of language from Paul Broca and Carl Wernicke, and further work by Gustav Fritsch, Eduard Hitzig, and David Ferrier on localization of the motor cortex (Robison et al., 2013). The study of patients with traumatic brain injuries was also essential, the most famous example being Phineas Gage, who sustained a penetrating injury to his dominant frontal lobe and subsequently developed aggressive and impulsive behavior (Damasio et al., 1994).

The Advent of Modern Psychosurgery
Inspired by these findings, Swiss psychiatrist Gottlieb Burckhardt performed the first psychosurgical treatments in 1888. These procedures, conducted on six patients with mental illnesses, consisted of topectomies to excise brain regions, thought to be responsible for aggression, agitation or hallucinations (Stone, 2001); Burckhardt was subsequently ostracized from the medical community. The research and practice of psychosurgery again entered into dormancy and was largely quiescent until 1935, when John Fulton and Carlyle Jacobsen described the role of the frontal lobe in short term memory, anxiety and aggression in a primate model (Fulton and Jacobson, 1935). In attendance of this lecture was Egas Moniz, a Portuguese neurologist who, together with the neurosurgeon Almeida Lima, developed the prefrontal leucotomy procedure for the treatment of psychiatric patients with prominent depression, anxiety or aggression (Feldman and Goodrich, 2001). This procedure initially consisted of alcohol injections into the prefrontal white matter, and Moniz eventually developed the leucotome, an instrument that could be introduced through frontal burr holes for precision lesioning (Moniz, 1994).

The Prefrontal Lobotomy and Enduring Controversy
This work of Moniz and Lima inspired the Americans Walther Freeman, a neurologist, and James Watts, a neurosurgeon, who further refined the prefrontal leucotomy technique to allow for larger lesions that presumably disrupted widespread brain networks associated with affective processing. The precision leucotome was subsequently developed, and procedural modifications were introduced to allow for more accurate lesion localization using anatomical landmarks (Heller et al., 2008). This refined procedure was named the prefrontal lobotomy: minimal lobotomies were predominantly performed for the treatment of affective symptomatology, whereas radical lobotomies were for schizophrenic patients or those with refractory symptoms (Feldman and Goodrich, 2001). Freeman and Watts performed hundreds of procedures, and described relative success in the treatment of depression, agitation and aggression, although they cautioned its use should be limited to those patients with severe and refractory symptoms (Freeman and Watts, 1937). Unfortunately, this led to indiscriminate use of the lobotomy, partially in response to the lack of conventional treatments for mental illness, and also as a potential measure of reducing asylum overcrowding (Heller et al., 2008).
Freeman became emboldened by his initial success, and co-opted a transorbital technique developed by the Italian psychiatrist Amarro Fiamberti. This involved the insertion an instrument resembling an ice-pick, the orbitoclast, through the orbit into frontal lobe white matter, which allowed for lesioning via a sweeping motion in the coronal plane (Pressman, 1988). This technique was much quicker than the standard lobotomy and only required electroshock therapy for sedation; as a result, Freeman did not require an anesthetist, surgeon, or even proper sterile technique. In response, Watts and the neurosurgical community distanced themselves from this practice, and Freeman went on to perform thousands of procedures across America (Heller et al., 2008). The transorbital lobotomy was widely and eagerly adopted throughout America and parts of Europe, with tens of thousands of procedures performed across nearly two decades (Braslow, 1999). To illustrate the sheer popularity within the scientific community for psychosurgery at the time, the 1949 Nobel Prize in medicine or physiology was awarded in-part to Moniz for the leucotomy (Tierney, 2000).

Stereotactic Neurosurgery and the Advent of "Minimalism"
Eventually, both professional and public opinion turned against the lobotomy as the associated morbidity and mortality became more apparent (Heller et al., 2008). A number of patients also developed a "post-leucotomy syndrome, " characterized by apathy, emotional blunting, and disinhibition. For these reasons, the medical community sought more scientifically rigorous surgical approaches focused on hypothesis-driven targeting with less aggressive lesioning and resections. The neurosurgeon William Scoville was an early proponent of these principles and performed selective undercutting of the orbitofrontal cortex (Scoville, 1949). Further improvements were realized with advancements in stereotactic neurosurgery, which superseded open and closed lesioning techniques.
The earliest stereotactic apparatus had previously been developed by Sir Victor Horsley and Robert Clark for animal experimentation (Clarke and Horsley, 2007), and its clinical translation was described by Ernest Spiegel and Henry Wycis (Gildenberg, 2001). Using x-ray ventriculography, the dorsomedial nucleus of the thalamus was targeted to treat agitation and psychosis (Spiegel et al., 1947). Further advances in psychosurgical stereotaxis led to the development of the stereotactic cingulotomy, capsulotomy, subcaudate tractotomy, and limbic leucotomy; rather than removing the entire frontal lobe to produce symptom relief, targeted lesions of white matter pathways and gray matter structures were found to be comparable in effect with fewer side effects (Lapidus et al., 2013). These procedures continue to be used for refractory psychiatric illness in contemporary practice. Additional targets, including the amygdala and hypothalamus, have also been described but are less commonly used.
The cingulotomy has been used for a variety of disorders, and involves lesioning of the cingulate gyrus and adjacent white matter fibers of cingulum bundle. In 1952, Sir Hugh Cairns described this technique for symptoms of anxiety and obsession (Whitty et al., 1952). The cingulotomy has since been used in the treatment of obsessive-compulsive disorder (OCD), treatment-resistant depression (TRD), and was similarly found to be effective for individuals suffering from chronic, intractable pain (Ballantine et al., 1967). The cingulotomy remains the most commonly performed psychosurgical procedure in North America (Lapidus et al., 2013), and its therapeutic effects are in-part believed to involve disruption of cortico-striato-thalamocortical circuitry.
Anterior capsulotomy has also been found to be effective for OCD, likely due to the disruption of frontothalamic fiber connections. It can be performed with radiofrequency coagulation, as originally described by Jean Talairach, or via gamma knife radiosurgery as described by Lars Leksell (Laitinen, 2001). Subcaudate tractotomy disrupts the orbitofrontal cortex from its thalamic and limbic connections, and its early use can be traced to Scoville's orbitofrontal corticectomy (Scoville, 1949). Its adaptation into a stereotactic procedure can be attributed to Geoffrey Knight (Knight, 1965), and was initially performed via bilateral insertion of yttrium rods to cause radiation necrosis before transitioning to radiofrequency coagulation (Lapidus et al., 2013). Subcaudate tractotomy has been described for TRD and OCD, although cingulotomy and capsulotomy have largely surpassed its use in contemporary medicine (Lapidus et al., 2013). Lastly, the limbic leucotomy combines cingulotomy with tractotomy to target both frontothalamic and cingulate circuitry. First described by Desmond Kelly andAlan Richardson in 1973 (Kelly et al., 1973), it is likewise used for TRD and OCD. Ultimately, these techniques share a common theme of wide-spread neural network disruption.

THE FALL OF PSYCHOSURGERY AND THE RISE OF PHARMACOLOGY
The decline of psychosurgery in the 1950s can be attributed to numerous factors. The initial enthusiasm for frontal lobotomies was driven by widespread social acceptance and Freeman's deft hand for advocacy and marketing. However, this also resulted in a laissez-faire attitude toward psychosurgery, with procedures being performed by non-neurosurgeons in inappropriate settings. This triggered professional criticism regarding the significant and largely underreported adverse events (Hoffman, 1949), and the lack of objectivity and scientific rigor. Furthermore, the public became aware of the undesirable consequences of lobotomies. Social attitudes were shaped by negative depictions in literature and film, including notable examples such as One Flew Over the Cuckoo's Nest. It also became apparent that some institutionalized or incapacitated patients were lobotomized without informed consent, and procedures may have been performed on prisoners to address dysfunctional behavior as opposed to mental illness (Feldman and Goodrich, 2001). The controversy surrounding psychosurgery was instrumental in developing modern standards for research and ethics, with the passing of the National Research Act of 1974 and subsequent publication of the Belmont Report, which outlined the principles of informed consent and guidelines regarding the performance of medical procedures and research.
However, it was the evolution of pharmacotherapy that finally turned the tide against psychosurgery, particularly with the advent of lithium (Cade, 1949) and chlorpromazine (Lopez-Munoz et al., 2005) therapy. In particular, the use of chlorpromazine in the early 1950s provided an effective non-surgical treatment for severe aggression and psychosis. As a result, psychosurgery fell out of favor and psychopharmacology thrived as brisk sales prompted investment and further innovation. The development and approval for other antipsychotic and antidepressant agents soon followed, with evidence that medical therapy was more effective, safer, and cheaper than psychosurgery.

Modern Psychopharmacological Treatments and Limitations to Medications
The standard first-line treatments for psychiatric disorders include medications and psychotherapy. Commonly used pharmacological agents are outlined in Table 1 (Gitlin, 2006;Marazziti and Consoli, 2010;Singer, 2010;Gao et al., 2015;Qaseem et al., 2016). Interestingly, there is significant overlap among psychiatric disorders and the pharmacological agents used in treatment, suggesting that these disorders likely involve widespread and shared circuitry and/or neurotransmitters. However, treatment resistance or failure is common, necessitating alternative strategies. The first strategy is commonly to optimize medication dosing or scheduling, or switching to another therapeutic class. TRD, for example, refers to an inadequate response to at least one antidepressant trial, and can occur in up to half of patients (Fava, 2003). Even with optimized multi-modal therapy, up to 30% of patients with major depression are non-responders (Rush et al., 2006). Similarly, a significant proportion of patients with OCD (Greenberg et al., 2003), bipolar disorder (Geddes and Miklowitz, 2013), and malignant Tourette syndrome (Cheung et al., 2007) fail to respond to standard therapy, again suggesting that the underlying pathophysiology is likely complex and involves multiple networks. These treatment-resistant patients often have more severe symptoms, and higher rates of morbidity and mortality. As such, new therapeutic strategies are desperately required.

OTHER NEUROMODULATION THERAPIES
Like psychosurgery, electroconvulsive therapy (ECT) was introduced before the advent of psychopharmacology. Despite its documented efficacy in the treatment of psychiatric disorders, its side effects on cognition and memory and misuse/abuse by some practitioners dramatically decreased its use before the 1980s. The realization of the imperfect effects of psychotropics "resuscitated" the use of ECT, although it is still considered under-utilized, primarily due to misconceptions regarding treatment (Bewernick and Schlaepfer, 2015). Although the exact mechanism is not understood, the induction of a generalized seizure and subsequent post-ictal suppression is effective for TRD (Sackeim et al., 2001) and bipolar disorder (Perugi et al., 2017), and may also be beneficial for OCD (Fontenelle et al., 2015).
Additional non-invasive therapies include magnetic seizure therapy, which focally induces the superficial cortex to produce seizures, repetitive transcranial magnetic stimulation, involving the pulse application of magnetic stimuli to alter cortical excitability, and vagal nerve stimulation, which sends electrical impulses to various brain regions via the solitary nucleus (Bewernick and Schlaepfer, 2015). The role for these alternative treatments remains unclear, as ECT is generally considered clinically superior (Dierckx et al., 2012;Berlim et al., 2013). However, the remission rates for TRD with ECT are reportedly from 35 to 53% (Schoeyen et al., 2015). Clearly, a significant number of patients will not respond to pharmacological and non-invasive neuromodulation treatments, suggesting that more invasive surgical interventions like deep brain stimulation (DBS) can be considered as a viable option for patients with psychiatric disorders who fail "all" available treatments.
Although DBS has shown efficacy in patients with treatmentresistant major depression and bipolar depression (Holtzheimer et al., 2012), the investigation of DBS for such diseases at TRD in larger scale randomized trials has been less successful (Dougherty et al., 2015;Holtzheimer et al., 2017). This may in part be due to the unclear determination of targets that would effectively produce the same results seen with lesioning techniques. Traditionally, DBS targets have been selected based on neuroimaging findings and historical results of lesioning procedures, though clearly further hypothesis-directed targeting studies are needed.  Anxiety, depression, hypotension, insomnia, nausea, sedation * specific drug interactions not listed. SSRI, selective serotonin reuptake inhibitor; SNRI, serotonin norepinephrine reuptake inhibitor; NDRI, norepinephrine dopamine reuptake inhibitor; NaSSA, noradrenergic and specific serotonergic antidepressant; SM, serotonin modulator; TCA, tricyclic antidepressant; MAOI, monoamine oxidase inhibitor; ATA, atypical antipsychotic; AC, anticonvulsant; DA, dopamine agoinst; DRI, dopamine reuptake inhibitor; TA, typical antipsychotic; A2A, alpha II adrenergic agonist; MAUI, monoamine uptake inhibitor.

FUTURE DIRECTIONS
There are numerous limitations to the current understanding and treatment of psychiatric disorders. These are complex disease states which cannot be localized to distinct anatomical correlates, and are influenced by cognitive, behavioral and emotional factors. As such, the contemporary understanding of psychiatric disease may be attributable to a dysfunction of networks rather than an isolated structural or neurotransmitter abnormality. This interpretation can be traced to seminal works including the description of the Papez and cortico-striato-thalamocortical circuits, which postulate that functional subdivisions influence various facets of behavior (Alexander et al., 1986;Papez, 1995).
If these diseases are truly abnormalities in wide-spread white matter connections and brain circuitry, then a better understanding of such pathways in the human brain is essential. While data from animal studies remains the gold-standard for the evaluation of white matter fiber tracts, these models may not accurately translate to the human correlate with such complex behavioral and cognitive processing. Moreover, although improvements in neuroimaging have been fundamental in advancing our understanding of the underlying neurobiology and circuitry involved in psychiatric disorders, these also have limitations in their ability to precisely and accurately allow the visualization of these connections. Innovations in neuro-imaging techniques will be necessary to help identify radiographic biomarkers indicative of a particular disease entity. Methodological innovations that may improve the accuracy of imaging technologies include advances in machine learning, optimization of signal prediction error, and the development of novel imaging algorithms (Maier-Hein et al., 2017). Furthermore, there is likely significant patient variability regarding this underlying circuitry. Riva-Posse et al. have previously demonstrated that patient responders to DBS for TRD shared similar fiber pathways that were being stimulated (Riva-Posse et al., 2014). Their subsequent prospective study utilizing probabilistic tractography resulted in a high patient response rate, suggesting an important role for targeting based on patientspecific data (Riva-Posse et al., 2018).
Advances in DBS technology and transition to a "closed loop" system may hold promise for the treatment of psychiatric disorders based on patient or disease-specific biomarkers. Conventional "open loop" stimulation requires the physician to adjust stimulation parameters based on subjective evaluations, whereas a closed-loop system adjusts stimulation parameters based on direct measurement of a neural biomarker (Lo and Widge, 2017). In Parkinson's disease, pathological beta band activity may represent an essential biomarker for adaptive stimulation (Moraud et al., 2018). However, biomarker identification in psychiatric disease states will prove to be more challenging due to clinical heterogeneity, and may be dependent on identifying commonalities among disease phenotypes (Bilge et al., 2018). The implantation of DBS electrodes allows for both acute (microelectrode) and chronic (macroelectrode) recordings to identify potential electrophysiological biomarkers in psychiatric disease states. Such recordings may provide insight into the electrophysiological correlates of acute symptomatology, and also long-term changes as a result of the disease process or chronic DBS therapy.
There is emerging evidence that local field potential activity from various brain regions may have distinct patterns in psychiatric disorders (Neumann et al., 2014;Merkl et al., 2016), and that invasive electrophysiology may allow for circuitry identification and modulation in selected patients . In a pre-clinical study of mice prone to binge-eating, Wu et al. observed that closed-loop stimulation of the nucleus accumbens disrupted rewarding behavior (Wu et al., 2018). Specifically, the identified biomarker was pathological delta band activity from local field potential recordings. Although promising, closed-loop stimulation has yet to be clinically validated in psychiatric disorders.
It has also been suggested that the lack of treatment success of DBS for psychiatric disorders in larger clinical studies may be attributable to the clinical trial design itself (Dougherty, 2018). Whereas the efficacy of DBS for movement disorders can be reliably measured based on objective markers, the diagnostic criteria and interpretation of psychiatric disorders can be more variable and complex, and the clinical response may take longer to manifest. As such, clinical trial design should take into consideration different therapeutic endpoints and/or a longer trial period.
Ultimately, the contemporary use of neuromodulation for mental illness will need to focus on a hypothesisdriven approach to target selection and will likely incorporate multiple targeting methods using novel technologies to produce effective results (Riva-Posse et al., 2014, 2018. Patient heterogeneity should not be undervalued, and an individualized approach may also be required, be it symptom-directed, patient-directed, or both (Holtzheimer and Mayberg, 2011).

CONCLUSION
The modern resurgence of surgical therapies for psychiatric disorders must respect the lessons learned from the early and mid-twentieth century, particularly the ethical and social ramifications of poor scientific rigor. The treatment of psychiatric patients continues to be challenging, especially in those patients deemed incapable of making informed decisions. For these reasons, the practice of psychosurgery should embrace a multidisciplinary approach to diagnosis and management. The future of psychiatric treatments will undoubtedly involve concurrent developments in the fields of behavioral neuroscience, neuroimaging, psychopharmacology, and neurosurgery.

AUTHOR CONTRIBUTIONS
MS and JS participated in manuscript conceptualization, design, and preparation. MS, EH, KG, JM, and JS provided critical revisions of the manuscript and reviewed the submitted version. KG, JM, and JS provided administrative support.

FUNDING
This work was supported by the National Institutes of Health Grant 2KL2TR000440.