Serendipity and Observations in Functional Neurosurgery: From James Parkinson’s Stroke to Hamani’s & Lozano’s Flashbacks Open Access

“You see but you do not observe”

Sir Arthur Conan Doyle: A Scandal in Bohemia, 1891

Serendipity has many definitions: “the fact of finding interesting or valuable things by chance”; “finding valuable or agreeable things not sought for”; an “unplanned fortunate discovery”; or a “happy discovery by accident.” For André Breton, a pupil of Joseph Babinski and the father of surrealism, “Chance” (“Le Hasard” in French) is “la rencontre entre une causalité externe et une finalité interne” (the meeting of an external causality with an internal finality) [1]. Hence, for “chance” to be able to lead to serendipity, one needs to be both observant and prepared.

Serendipity, as well as fortunate observations by prepared minds, has been responsible for many inventions and discoveries in technology (for example, the internet, Velcro, superglue, and the microwave oven). Serendipity has also been responsible for momentous medical discoveries such as the discovery of penicillin [2], the use of electro-convulsive therapy to treat depression [3, 4], carbamazepine for the treatment of trigeminal neuralgia [5], sildenafil for the treatment of erectile dysfunction [6], amantadine in the treatment of Parkinson’s disease (PD) [7], and even the discovery of the role of 3,4-dihydroxyphenethylamine, dopamine, in the pathogenesis of PD [8, 9].

In functional neurosurgery, it is often assumed that many of the chosen targets for ablation or stimulation stemmed from research on animals and/or clinico-pathological studies. Could it be, however, that pure serendipity played a significant role in the discoveries of “new” brain targets, new indications, and new surgical procedures? The aim of this review is to critically examine this possibility.

Scholarly publications, including peer-reviewed papers, books, and book chapters relevant to the history of functional neurosurgery for movement and psychiatric disorders, were reviewed, both concerning the pre-stereotactic as well as the stereotactic era. As far as possible, original documents were retrieved and scrutinized with respect to indications for surgery, the surgical methods, and the brain targets, in view of determining whether serendipitous discoveries and observations contributed to the proposal or establishment of a variety of functional neurosurgical procedures.

In his pamphlet “An Essay on the Shaking Palsy” published in 1817, Parkinson [10] described in 6 patients the clinical features that came to carry his name. In patient no 6, he noticed that the symptoms of tremor on the right side had disappeared after the patient suffered a stroke: “During the time of their having remained in this state, neither the arm nor the leg of the paralytic side was in the least affected with the tremulous agitation” [10]. This astute observation by the surgeon apothecary encouraged neurosurgeons during the first half of the 20th century to attempt treatment of Parkinsonian tremor by surgical interventions directed at the upper motor neuron from the motor cortex to the spinal cord [11]: Bucy and Case [12] and Klemme [13] performed subpial resection of the premotor cortex area 6 and the motor cortex area 4 in patients with tremor; Browder targeted the anterior internal capsule, extending his resections caudally towards the motor part until the tremor resolved [14]; Lateral pyramidal tractotomies were performed at the level of the upper spinal cord by Putnam [15], Ebin [16] and Oliver [17, 18], while Walker [19] and Guiot and Pecker [20] championed mesencephalic pedunculotomy. All these procedures caused hemiparesis and it was believed that the reduction of tremors was proportional to the degree of paralysis. Sometimes Parkinsonian rigidity was replaced by postoperative spasticity. All this was in line with the statement of Bucy and Case [12] that “neurosurgical alleviation of tremor was dependent upon production of paralysis or damage to the pyramidal tract or to corticospinal pathways.”

The basal ganglia, which is the modern name for what was once called the extra-pyramidal system, was described by Walter Dandy as “the center of consciousness” and therefore a “no-man’s land” for the neurosurgeon [21]. Meyers [22], a young neurosurgeon from Iowa, was the first to challenge this view by operating on the caudate nucleus of a patient with Parkinson’s disease (PD). In 1939, he resected the head of the caudate in 1 patient, leading to sustained relief of tremor and rigidity [22]. He subsequently proceeded to extend the anatomical scope of his resections, reaching the level of the internal pallidum and the ansa lenticularis [23]. According to Gabriel and Nashold [24], Meyers’ decision to resect the head of the caudate arose from a chance observation by Browder. During a frontal lobotomy for depression in a patient with Parkinsonism, the incision was extended inadvertently to the caudate nucleus. When the patient awoke, Browder noticed that the shaking had abated [24].

One of the American pioneers of surgery for movement disorders was Cooper [25], who in 1951, while attempting a pedunculotomy on a PD patient, injured the anterior choroidal artery and had to ligate it. The surgery was terminated before the pedunculotomy was carried out [26]. The patient awoke with no neurological deficit and without a tremor. This led Cooper [27] to propose surgical occlusion of that artery as a treatment for Parkinsonism. Subsequent autopsies of patients who had undergone the procedure revealed degeneration of the medial globus pallidus. This pathological finding led Cooper to direct his attention to lesioning the medial globus pallidus using an unusual temporal approach in which he directed the probe upwards and inwards towards the pallidum, a procedure which would eventually lead to yet another serendipitous discovery.

The first use of stereotactic surgery in humans by its pioneers Spiegel and Wycis was to alleviate severe mental illness. They produced a small focused lesion in the dorsomedial nucleus of the thalamus [28] based on anatomical observations that after prefrontal leucotomy, a retrograde Wallerian degeneration affected the thalamus, especially its dorsomedial nucleus [29, 30]. The degeneration involved fronto-thalamic fibers traveling in the anterior arm of the internal capsule and Talairach in France favored this site, labeling the procedure “lobotomie frontale limitée” [31] which later was referred to in the Anglo-Saxon literature as anterior capsulotomy.

During a rostral ventral anterior cingulotomy, Laitinen performed high-frequency stimulation of the genu of the corpus callosum, which resulted in “a sudden strong feeling of inner wellbeing and relaxation of the whole body of the patient,” and he reported that “The response was reproduced several times.” He then lesioned this area, which resulted in the disappearance of “all anxiety, fears, and obsessions” in the patient. This led Laitinen to propose a new procedure that he labeled “anterior mesoloviotomy” (mesolovium is the Greek name for the corpus callosum) to treat “intractable anxiety, fears, and tension of neurotic, schizophrenic, or epileptic origin” [32].

Russel Meyers’ open surgical procedures on the basal ganglia, and especially on the pallidum and its outflow, the ansa lenticularis, demonstrated for the first time that it was possible to treat tremor without damaging the corticospinal pyramidal tract, but at the expense of severe morbidity and high mortality. Cooper then showed that it was degeneration of the internal pallidum following ligation of the anterior choroid artery that was responsible for the decrease in the motor symptoms of Parkinson’s disease. The pallidum and its outflow tracts, therefore, became a logical lesional target for the newly developed stereotactic techniques [26, 33].

In Germany, ventrolateral thalamotomy was developed based on the anatomo-pathological studies of Hassler and Riechert [34, 35], whereas in North America, a patient whose tremor had been abolished by a pallidotomy carried out by Cooper using his lower temporal approach was found at autopsy after a fatal car accident to have a lesion not in the pallidum but in the ventrolateral thalamus [26, 36]. As a consequence, by the late 1950s, most neurosurgeons started to target the ventrolateral thalamus for Parkinson’s disease, including some who targeted the subthalamic area, in which it was empirically discovered that lesions were very efficient in alleviating tremor [37‒39].

However, Lars Leksell in Lund, Sweden, continued to perform pallidotomies targeting various areas of the globus pallidus internus (GPi). Among the 81 patients of Leksell, analyzed by neurologist Svennilson et al. [40] in their seminal paper from 1960 [41], the best results occurred in the last 20 patients of the series where lesions were put in the posteroventral part of the internal pallidum. Leksell had “discovered” by chance that lesions in this region of the pallidum were highly effective for treating the symptoms of Parkinson’s disease before the existence of the “basal ganglia model” and long before anybody knew that this region harbored the sensorimotor circuitry of the globus pallidus.

It was Svennilson’s paper as well as personal discussions with Leksell that prompted Laitinen et al. [42] to revive “Leksell’s posteroventral pallidotomy,” the results of which were published in 1992. Laitinen began this approach in January 1985 [43], at a time when the classical Alexander-DeLong model of segregated basal ganglia circuitries had still not been described [44]. The fortunate observation that levodopa induced dyskinesias and painful dystonia (that did not exist in the 1950s when Leksell performed his operations) were highly relieved by posteroventral pallidotomy (PVP) paved the way for surgical treatment of non-Parkinsonian dystonias with PVP [45], and later with posteroventral pallidal DBS [46‒48].

DBS as an investigational, then as a treatment method of brain diseases has its roots in psychiatry [49, 50]. An early proponent was Robert Heath at Tulane University in New Orleans. In 1950, he started a DBS program aiming at treating patients with schizophrenia and chronic pain, and implanted electrodes in various subcortical areas of the brain involved in emotions such as the head of the caudate, the amygdala, the septal area, the thalamus, and the hypothalamus [51]. Later, he became interested in the potential of DBS for enhancement of pleasure. This shift occurred as a consequence of two observations: first, the serendipitous discovery in 1954 of the positive rewarding effect of stimulation of the septal area in rodents [52]; the second was Heath’s own observations of the behavior of schizophrenic patients implanted with brain electrodes and who could not stop self-stimulating themselves. Heath wrote: “In hindsight, however, the most interesting observation was that, for a small group of patients, electrical brain stimulation was pleasurable. Two patients described the experience as “pleasant.” Two others were described as “jovial” or “euphoric” after brain stimulation. In all four cases, the area that was stimulated was, reportedly, the septal region” [53‒55].

During a ventral intermedius (VIM) thalamotomy for tremor, Alim Louis Benabid experimented with high-frequency stimulation of the ventrolateral thalamus and observed a complete cessation of tremor as long as the stimulation was on, and without deleterious effects for the patient [56, 57]. This chance observation, detailed by Benabid in his recently published book “Le Hasard et le Possible” (Chance and the Possible) [58], established a new therapeutic approach that would extend far beyond the VIM of the thalamus and Parkinson’s disease.

In 2002, in the Lancet, a report appeared describing two people with Parkinson’s disease and obsessional compulsive disorder whose Parkinson’s disease had been treated with DBS of the subthalamic nucleus (STN) [59]. In addition to improvement in Parkinsonian symptoms, the patients’ obsessions had reduced markedly. Postoperative imaging showed that the DBS leads had been placed in the vicinity of the anteromedial “limbic” part of the STN. This chance observation led to a randomized controlled trial (RCT) of DBS in the limbic area of STN for treatment of OCD [60], followed by few other reports in which stimulation in the anteromedial STN was shown to be beneficial for severe OCD [61, 62].

A similar serendipitous discovery was published by Coenen et al. [63]. A patient with PD who had DBS in the STN exhibited bouts of hypomania upon stimulation. Detailed analysis of the anatomical location of the electrode combined with tractography imaging and analysis of the volume of brain tissue encompassed by the electric current showed that the stimulated contact provoking hypomania was very close to the anterior limbic STN and to fibers of the superolateral medial forebrain bundle (MFB). This led the authors to use the MFB area as a target for DBS in patients with severe refractory depression [64] and later also in patients with severe refractory OCD [65]. The debate on whether the effect of DBS on the psychiatric symptom emanates from the anteromedial STN or from the MFB remains unresolved [65, 66].

Mayberg et al. [67] seminal paper from 2005 described the effect of DBS of the cingulum area 25 in 6 patients with major depressive disorder. The choice of that brain target was based on their previous observation on functional imaging studies “that the subgenual cingulate region (Brodmann area 25) is metabolically overactive in treatment-resistant depression,” and that there was “Consistent involvement of the subgenual cingulate (Cg25) in both acute sadness and antidepressant treatment effects.”

The first report on DBS for Gilles de la Tourette (GTS) was published by Vandewalle et al. [68] in The Lancet. The authors targeted medial thalamic nuclei based on an old publication by Hassler and Dieckmann [69] on ablative stereotactic surgery for that condition. van der Linden et al. [70] in Belgium and Diederich et al. [71] in Austria then reported successful treatment of tics by DBS of the posteroventral internal pallidum, i.e., the same Leksell-Laitinen target used for PD and that had also recently been used successfully for dystonia. The rationale for the authors to choose the posteroventral globus pallidus internus (GPi) for DBS in GTS was the assumption that tics, like dystonia, are mediated by the main motor output center of the basal ganglia, which is the sensorimotor GPi. Diederich et al. [71] followed their patient for 14 months and wrote: “Similarly to DBS in dystonic patients, the improvement was noticeable only several weeks after DBS implantation but progressed during the entire observation period…”

In a patient with depression and anorexia nervosa who underwent DBS of the subgenual cingulum, it was observed that the eating disorder improved, while the depression did not [72]. This led to the first trial of DBS for anorexia targeting the CG25 [73].

In OCD patients treated with DBS of the nucleus accumbens, it was observed that those individuals who were heavy smokers or heavy drinkers, craving for nicotine, and alcohol decreased [74, 75]. This led on to further so far unconclusive but “promising,” trials aiming at treating alcohol as well as other drug addictions [76‒79].

Perhaps the most remarkable chance discovery during DBS happened in 2008 when the Toronto group attempted treatment of obesity by hypothalamic DBS. During intraoperative stimulation, the patient reported memory flashbacks on several occasions. Analysis revealed that the electrodes lay very close to the fornix [80]. This chance finding [81] sparked interest in treating Alzheimer dementia with DBS, with a phase 1 RCT [82], a phase 2 RCT [83, 84] as well as an ongoing multicenter RCT for people with mild Alzheimer aged 65 years or more.

A similar observation of occurrence of memory flashbacks was noted by another group during DBS of the ventral capsule for OCD in 2006 [85]. The authors wrote: “1 patient, after an increase in amplitude at the distal contact (0), developed brief memory experiences for events surrounding the surgery itself. These recurred several times a day over several days and ceased when she reported them and parameters were changed.” A further similar observation was reported by the same group in a paper in 2010 [86], in which the authors wrote: “Another patient developed brief memory experiences (“flashbacks”), recurring several times a day over several days, which ceased after parameters were changed.” These observations, which are very similar to Hamani’s and Lozano’s observations mentioned above, did not lead to further action and exploration of the anatomical rationale for the reported memory “flashbacks.”

Thus, it is evident that, ever since its inception, functional neurosurgery has owed much to observation and serendipity. Most notable in the pre-stereotactic era were the heroic surgical forays into the basal ganglia in the treatment of PD, based on an “accidental” discovery following a frontal lobotomy in a patient with psychiatric illness and concommittant PD. In psychosurgery it was autopsy studies revealing the pathways along which retrograde degeneration traveled after lobotomy, that contributed to the first applications of, and first brain “targets” for, the newborn human stereotactic technique.

Serendipitous observations, some of which led to discoveries, continued to pop up all along the stereotactic lesional era, and even more during the modern DBS era, in part thanks to technological advances and advances in brain imaging. This, in turn, has led to consideration of potential new brain targets to treat various behavioral, neurological, and psychiatric disorders. As shown in Table 1, it seems that discoveries in functional neurosurgery occurred along different serendipitous journeys: clinical observations in general, intraoperative observations, observations on imaging studies, and pathological observations on autopsy.

The evolution of modern DBS started by chronic stimulation of previously lesioned brain targets: the VIM for tremor, posteroventral pallidum for PD, anterior internal capsule for OCD and depression, central thalamic nuclei (center median, ventral oral anterior) for Gilles de la Tourette syndrome, subthalamic area including zona incerta – but excluding the STN proper – for tremor. None of these DBS targets, not even the first modern era DBS of Benabid et al. [87], did result from animal experiment. Only three of the current human applications of DBS stemmed from prior work on animal: the lesioning of the STN in primate models of PD [88, 89] led to high-frequency DBS of the STN in human PD; the low-frequency stimulation of the pedunculopontine nucleus in monkey [90] led to attempts of low-frequency DBS in PD patients with freezing of gait; and the numerous works on DBS in rodent models of post-traumatic stress disorder (PTSD) led to a cautious trial of DBS of the basolateral amygdala in patients with PTSD [91, 92]. In fact, some of the serendipitous discoveries during human clinical DBS have led researchers in the opposite direction, that is, from bedside to bench. One example is the confirmation in rat models of the beneficial action of fornix DBS on memory [93]; another example is the demonstration that DBS in the medial forebrain bundle may be effective in rodent models of depression [94].

While some serendipitous observations have led to shifts in neurosurgical practice, others did not. In 2004, The pioneers of DBS for Gilles de la Tourette observed that thalamic high-frequency stimulation (HFS) had an influence on erection in two of their first 3 patients. They concluded that “the midline and intralaminar thalamic nuclei play an important role in penile erection and that erection can be modulated by HFS of this region” [95]. Despite further studies from this same group on brain networks for erection [96‒98], nothing clinically valuable resulted from these observations.

DBS of the pedunculopontine nucleus (PPN) for treatment of freezing of gait has also generated serendipitous observations: In 2009, an Italian group reported that stimulation of the PPN showed a “trend towards reduction of ungrammatical errors” and “provide the rationale for further investigation of the role of the PPN in processing linguistic grammar” [99]. There were no further follow-ups to this discovery.

In 2010, the Grenoble group noticed that during low-frequency stimulation of the PPN area patients exhibited increased alertness, whereas high-frequency stimulation induced sleep [100]. The authors concluded that their observations have the potential to benefit patients who suffer from sleep disorders. There were no further explorations of this serendipitous discovery and as one of the authors put it later, the risk of implanting DBS leads in this eloquent area for such an indication largely outweighs the potential beneficial effect on the sleep disorder [58].

In 2007, the Oxford group reported intraoperative modulation of blood pressure in a patient with chronic pain undergoing DBS of the periventricular/periaqueductal gray (PAG/PVG) area [101]. On the basis of this single observation, it was concluded that “hypertension could be effectively treated with electrical stimulation of the midbrain.” However, the effect on blood pressure on that patient was inconsistent and was reversed upon evaluation 5 years later [102]. The same group had gone on to study on purpose autonomic “side effects” of DBS affecting blood pressure, respiration, micturition, etc. in attempts to verify their consistency in view of possible new indications for DBS, and in 2012 it was stated that DBS may in the future be able to treat hypertension, orthostatic hypotension, asthma, chronic obstructive pulmonary disease, obstructive sleep apnea, bladder detrusor malfunction and some other autonomic disturbances [103]. Sometimes a seemingly serendipitous “discovery” may be just a “side effect” and will thus loose its definition as being an “unplanned fortunate discovery,” once it becomes neither “unplanned” nor, alas, “fortunate.”

A review of functional neurosurgery has revealed the role of serendipity in many discoveries of “new” indications and “new” brain targets for lesioning and/or for DBS. The advent of modern DBS more than 30 years ago has accelerated this trend. In a recently published paper titled “Where are we with deep brain stimulation? A review of scientific publications and ongoing research” [104], the authors list more than 50 clinical disorders on which DBS has been applied and over 30 brain regions that have been targeted with DBS. A non-negligible number of the listed indications for DBS and brain regions targeted are indeed the results of observations and serendipity. As Louis Pasteur put it in 1854 [105]: “In the fields of observations chance favors only the prepared mind.”

Marwan Hariz has received fees and travel expenses from Boston Scientific for speaking at meetings. Andrew J. Lees is funded by the Reta Lila Weston Institute of Neurological Studies, Institute of Neurology, University College London, and reports consultancies from Britannia Pharmaceuticals and BIAL. He also reports grants and/or research support from the Frances and Renee Hock Fund and honoraria from Britannia Pharmaceuticals, Profile Pharma, UCB, Roche, BIAL, STADA, NordicInfu Care, and NeuroDerm. Yulia Blomstedt has nothing to declare. Patric Blomstedt is consultant for Medtronic, Abbott, and Boston Scientific and shareholder in Mithridaticum

This article did not receive any funding.

Marwan Hariz contributed to study design, study conduct, data collection, analysis and interpretation, first manuscript drafting, revision, and approval. Andrew J. Lees contributed to data collection, analysis and interpretation, manuscript drafting, revision, and approval. Yulia Blomstedt contributed to analysis and interpretation, revision and approval. Patric Blomstedt contributed to analysis and interpretation, data collection, revision, and approval.