The stereotaxic operations were carried out in two parts. A modified Bertrand stereotaxic instrument was used for implanting the depth electrodes. During the first operation, electrodes were introduced stereotaxically through a frontal burr hole usually situated 1 cm behind the coronal suture and 1 cm lateral to the sagittal suture. Depth electrodes were implanted in the side of the brain opposite the area of the pain. Multiple electrodes were implanted through the burr hole, and electrical stimulation was carried out in the weeks that followed. This was done while the patient was alert and cooperative. At a second operation, one or more H. F. lesions were made in the dorsolateral tegmentum at the level of the superior colliculus to interrupt the ascending pathways mediating painful sensations.

At the first stereotaxic operation the ventricular system was outlined with air and Pantopaque (3 cc). This was done by introducing a small flexible catheter through the lateral ventricle via one intraventricular foramen into the third ventricle. The posterior commissure and rostral end of the Sylvian aqueduct were always visualized by this technique, which was considered essential for the precise stereotaxic orientation of the electrodes into this region of the midbrain. The electrodes were aimed in the rostrocaudal direction to cross the anterior commissure-posterior commissure (AC-PC) line at an approximate angle of 65° to 70°. The point of entry for the electrodes into the dorsolateral portion of the midbrain was usually at the level of the posterior commissure and superior colliculus. The electrodes were also oriented in a lateromedial direction so as to enter the midbrain at an angle of 2° to 4° to the midsagittal plane through the center of the third ventricle or the aqueduct.

With this method of introduction the electrodes came to lie almost parallel to the long axis of the brain stem and passed in a rostrocaudal direction through the core of the dorsolateral and ventral tegmentum as far caudally as the inferior colliculus (Fig. 1). The electrodes introduced in this manner came to lie in the dorsolateral tegmentum at varying distances from the midsagittal plane. It was estimated that medial electrodes were lying within 2 mm of the center of the aqueduct while the most lateral electrodes were situated at distances from 5 to 12 mm from the midline. The electrodes in the central gray were nearly parallel to the Sylvian aqueduct.

X-ray films of the skull at the time of stereotaxic implantation of depth electrodes. Ventricular system outlined by oxygen and Pantopaque. The electrode introducer can be seen in A and B. The electrodes implanted in midbrain in C and D.

The lateral electrodes from 5 to 12 mm passed through the lateral tegmentum in the vicinity of spinothalamic and lateral spinothalamic tracts. The rostral contacts on these electrodes were situated in the region of the sensory thalamus, centre medianum, and pulvinar. Some electrodes were situated in the ventral tegmentum in or near the medial lemniscus and red nucleus. The position of the electrodes was estimated from several stereotaxic atlases of the brain.^12,15 The results of stimulation were gathered from approximately 34 electrodes situated in the diencephalon and mesencephalon (Fig. 2). No neurological deficits were produced by the implantations except that of a man (Case 5) whose causalgic pain was completely relieved after the introduction of three electrodes into the dorsolateral midbrain. He developed a mild contralateral analgesia to pinprick and reduced thermal sensation over the painful area.

Schematic sagittal view of the third ventricle and neighboring structures (adapted from stereotaxic of Schaltenbrand and Bailey¹²). Composite picture of electrode tracts at the time of implantation. A. Electrodes situated in the medial tegmentum and central gray in the zone 0 to 5 mm lateral from the midsagittal plane of the aqueduct. B. Electrodes located in lateral sagittal planes of the thalamus and mesencephalon. Electrodes with continuous lines located in a zone 5 to 10 mm from the midsagittal plane. The electrodes with the interrupted lines lie 10 to 12 mm from the midsagittal plane.

The details of the construction of the electrodes have already been described.⁵ They were multicontact stainless steel electrodes with six points for stimulation. Interelectrode distances varied from 2 to 3 mm depending on the type of electrode. The composite six-stranded electrode was 1 mm in diameter, and each single wire was coated with Teflon. They remained in situ in the brain for 2 to 3 weeks, and every week during the study their position was verified by skull x-ray. No changes were noted in their positions during the period of observation and stimulation.

Bipolar stimulation was done between each adjacent contact of the same electrode. A Grass square wave generator (S4) with an isolation unit was used for stimulation. Unidirectional square waves of 0.1 to 1.0 msec duration were delivered as single shocks or repetitive shocks ranging from 5, 25, 30, 60, 125 and 300 cps. The current flow was monitored at intervals on an oscilloscope. The electroencephalographic monitoring was done on a Grass 8 channel EEG (Model 6) from the adjacent non-stimulated contacts or from the adjacent electrodes or from electrodes placed on the scalp.

The patient was always allowed to recover from the surgery for a period of at least 1 week before the stimulation was started. During this interval the electrode resistances were measured each day. The electrical resistance of the contact points on this type of electrode ranged from 24,000 to 30,000 ohms. Standard scalp and depth EEG recordings were also done during this time to monitor the electrical activity of the brain.

The stimulation session was always discussed in detail with each patient so as to allay any fears, and to gain his confidence and cooperation in the observations. The patient was asked to give his own impression of his feelings and responses related to the stimulation. Every effort was made not to introduce leading or misleading questions. The patient was told he could stop the session of stimulation at any time. Stimulation was done in a quiet room with the patient sitting in a comfortable lounge chair. No medication was given at the time of the observations. During the initial stimulation sessions only weak stimuli were delivered to the brain and threshold values of the various responses were noted for each succession of contacts. As the sessions progressed, additional physiological recordings were made including measurements of the cardiovascular, pulmonary, autonomic, and ocular functions. Particular attention was paid to changes such as sweating (GSR measurements), skin temperature, changes in the color of the skin. Plethysmographic recordings were made in certain instances. A psychiatrist (W. P. Wilson) attended most of the sessions and noted the patient's over-all emotional responses to each stimulation. Special attention was also directed to the various facial expressions associated with the stimulation. Several medical observers were present at each of the sessions and detailed notes were kept plus cinematographic records, which were studied later.

The reporting of experiential, sensory, and emotional responses in persons who already suffered from intractable pains was difficult both for the patient and the observers. We made every effort to obtain objective answers from each patient, and without his interest, attention, and cooperation these studies would not have been possible.

The implantation of chronic electrodes in the brain of humans was not undertaken without careful thought and consideration as to the safety of the patient. The use of this technique was begun over 5 years ago, and the Duke Medical Center Committee on Human Experimentation has scrutinized the techniques and clinical results during this time. Each patient and his family were carefully and fully informed as to the risks involved, and it was stressed that the primary reason for the technique was to improve the localization of the therapeutic lesion for the improvement of the intractable pain in the individual patient. The patient was informed that he was free to stop the evaluation at any time. A team of three physicians and a nurse was responsible for the patient's evaluation and were present during the stimulation sessions.

The relief of pain by lesions in the midbrain requires precise localization of the lesion based on anatomical localization from roentgenograms and information gathered by stimulation of the ascending pathways mediating pain. Spiegel and Wycis¹⁶ emphasized this point that localization is dependent on the observations made by stimulating the midbrain. They carried this out during the surgical procedure when the patient was under a certain stress. Observations at such a time must be brief and often incomplete, thereby reducing the precision of the procedure.

The use of chronic implanted electrodes allows for a careful analysis of the effects of stimulation by the patient and observers. A second consideration in our technique was the possibility of predicting the therapeutic outcome of a lesion on the patient's spontaneous pain as gauged by the effects produced by the stimulation of pain pathways. All the patients were interviewed before and after removal of the electrodes and operation to determine whether there had been deleterious psychological effects. A further justification for these detailed studies was based on our desire to develop an improved stereotaxic technique for relief of these severe intractable pains in patients heretofore resistant to the current methods of treatment. We have followed the philosophy espoused by Penfield as a result of his long experience in the surgical treatment of the epileptic. He expressed it as follows: “No one who treats cases may escape criticism unless he is willing to study them with every means available, unless he recognizes that there are many different causes and each type requires appropriate therapy.”

In the alert human being, electrical stimulation of the dorsolateral mesencephalic tegmentum (Table 2) resulted in a series of complex sensory, motor, autonomic, and emotional responses. Only the details of the sensory responses will be reported here.

For the purposes of comparing the results of stimulation the dorsolateral tegmental area was divided into two separate zones. The “medial” zone was the area just lateral to the aqueduct extending from the midline laterally to a point 5 mm. This zone includes the periaqueduct gray, the tegmentum surrounding it. The “lateral” zone was the one between 5 to 12 mm from the midsagittal plane and included the region of the quintothalamic tract, its adjacent lateral tegmentum and the lateral spinothalamic tract.

The quality and kinds of sensations were not only influenced by the site of the stimulation but also by specific changes in the frequency and voltage of the current. The patient became quickly aware of the sensations even at low voltages and the level at which he became aware of a particular feeling did not seem to vary from day to day. He could identify the feelings such as “numbness,” “electric shock,” “burning,” “pain,” and he would also describe them as “weak,” “strong,” “intense,” “unpleasant,” or “fearful.” The exact adjective that he used to describe his feelings might vary from day to day, but with repeated stimulations at the same site he always had the same quality of feeling. The patient was always definite as to the referred location of the feeling. He might describe it as superficial “in the skin,” “deep in the bones,” “deep in the chest,” or “around the heart.” There seemed to be a difference between the quality of the sensations from stimulation of the sites in the region of the lateral spinothalamic tract as compared with sites in the region of the medial lemniscus. The feelings evoked from the lateral spinothalamic region were described as “brighter” or “sharper,” and a painful feeling was more easily elicited. On the other hand, the sensations evoked from the region of the medial lemniscus were less intense and they were described in terms such as a “numbness” or a “tingling” and, although a painful feeling could also be evoked from this region, higher intensities of stimulation were required and the painful feeling was never as intense.

The sites of stimulation in the midbrain were an important factor in determining the localization of the sensations as they were referred to the various portions of the body. The patient always referred the sensations elicited from the central gray region to the center portions of his body. In contrast to this he referred the sensations elicited from the spinothalamic and lemniscal regions to the contralateral parts of the body (Fig. 3). The ability of the patient to tolerate repeated stimulations was also dependent on the site activated in the brain. The patient's tolerance was lowest to the unpleasant feelings that originated from the medial region of the tegmentum but the patient could be stimulated repeatedly in the region of the spinothalamic and medial lemniscus even though he experienced an intense pain. The pain from the central gray region had a diffuse quality. This unpleasant feeling was always localized to the central portions of the body, as noted above, and centered in regions deep within the head, neck, chest, or abdomen. Certain of these sensations, which were localized to the head, were felt within the oral cavity involving the tongue, gums, and the teeth. One man complained of a metallic taste which was associated with his oral sensation. If the sensation was referred to the chest it was often felt in the region near the heart; however, we observed no changes in the patient's heart rate associated with the cardiac sensation. One woman complained of periumbilical and bladder sensations and she had an urge to void. If the intensity of the stimulation was increased, the patient also experienced an additional feeling described as fear which was so unpleasant that she was not willing to tolerate repeated stimulations (Fig. 4).

The figure on left shows area of referred sensation from lateral region of mesencephalon (crosshatched). The right figure shows midline reference of sensations from the central gray zone 0 to 5 mm from Sylvian aqueduct (dotted lines).

A coronal section of the dorsal mesencephalon at the level of the superior colliculus. The physiological responses elicited by stimulation of the dorsolateral mesencephalon.

The responses to the stimulations in the lateral regions of the dorsolateral tegmentum (5 to 10 mm from aqueduct) were of a different character from those elicited from the central gray region. The activation of these regions in or near the quintothalamic or spinothalamic tracts and lateral to them resulted in sensations referred by the patient to the contralateral portions of the body. Sensations referred to the opposite side of the face were elicited from sites in the midbrain medial to those sites from which sensations were referred into the arm or chest. The facial sensations were felt in the superficial cutaneous areas of the forehead, cheek, or jaw. When a facial sensation occurred it was found that by increasing the frequency or voltage of the stimulation the patient would note a spread of the area of referred sensation to include larger areas of the face, the shoulder, or the arm. There were few isolated instances when the sensation was referred also to the leg alone but most of the time it was in combinations with feelings in the chest or abdomen (Fig. 4).

Changes in the frequency of the stimulation resulted in differences in the kinds of feelings plus the appearance of the autonomic responses and facial expressions. Stimulations of the tegmental region near the medial lemniscus resulted in a tremor in the contralateral arm (6 cps) at lower frequencies from 2 to 10 cps. With an increase in the frequency above 10 cps, the patient complained of a sensation of movement in the arm. At higher frequencies (60 to 120 cps) the sensation of motion was replaced by a paresthesia which was with frequencies of 120 to 300 cps described as a “hot and painful sensation” (Fig. 5). This sequence of tremor, sensations of movement, followed by paresthesia and pain resulted from stimulating the region of the ventral tegmentum near the medial lemniscus. With stimulations of the dorsolateral tegmentum, a succession of sensations occurred. At low frequencies a tingling or numbness was felt, but above 60 cps the sensation was painful. No motor phenomena were noted from the region. Paresthesia were noted from thalamic stimulations but never pain. Two persons had a “feeling of movement” elicited from the region of the pulvinar. One had a phantom arm sensation with the sensation of motion referred into the phantom; the other patient had feeling but no movement was observed. The autonomic phenomena were also activated above 60 cps and were best observed at levels of 120 to 300 cps. The autonomic response never occurred singly but always in combination such as piloerection, blushing, or altered respirations.

Effect of frequency and voltage on the sensory responses. Sensory and motor responses from region of the ventral tegmentum and medial lemniscus at level of superior colliculus. The lower frequencies result in rhythmical tremor in contralateral arm. Higher frequencies result in no motor response but sensory phenomena. Unidirectional square waves (Grass stimulator). Pulse duration 1 msec.

A variety of painful and unpleasant sensations occurs in man and animals when the dorsolateral region of the mesencephalic tegmentum is electrically stimulated. This lateral portion of the midbrain contains both the direct and diffuse ascending sensory pathways involved in the mediation of noxious and painful sensations. The lateral spinothalamic tract in the mesencephalon has been the chief stereotactic target for the relief of intractable pains originating from the opposite face, arm, and chest. Spiegel, Wycis, and others have stimulated this region at the time of stereotaxic placement of therapeutic lesions.^2,9,11,16 They also noted that the subjects' painful responses elicited from the stimulation of the dorsolateral tegmentum were referred to the opposite face, arm, chest, abdomen, or leg. The reference of pain into the leg was the least common response, and it was assumed from this that the fibers transmitting in pulses from the leg had a more dorsal position in the lateral mesencephalon. Walker pointed this out in his anatomical studies of the midbrain.¹⁸

In our patients, few sensations were referred to the leg alone even from sites as far lateral as 12 mm from the aqueduct. Due to the orientation of our electrodes it was not possible to stimulate the most lateral and dorsal edge of the midbrain where most of the leg fibers appear to be located. Despite the fact that the stimulations were done acutely and at the time of the stereotaxic operation, Spiegel and others noted a definite topographic pattern of the referred pain as it was related to the various sites of the brain stimulation.^2,16 If the stimulation was done laterally in the vicinity of the quintothalamic or lateral spinothalamic tracts, the pain was referred to the contralateral region of the body such as the face, arm, chest, or abdomen. Those authors noted that stimulations done more medially caused the patient to complain of vague bilateral sensations experienced as a “choking” feeling referred into the chest.¹⁶ Spiegel inferred from these observations that noxious sensations from both sides of the body were traversing this medial region of the mesencephalon. These sensations were thought to be mediated via the diffuse spino-reticulo-thalamic system.^1,6,16

Sensations referred to the regions of the face were elicited from wider areas of the midbrain than have been noted previously These facial feelings occurred at the levels of the superior and inferior colliculus as close as 2 mm from the midline, probably from an area in or near the central gray, and were also noted as far laterally as 5 to 12 mm in the lateral tegmentum. This also suggested to us that there were multiple ascending pathways related to various regions of the face. The medial portion of the midbrain near the aqueduct has fiber tracts related to the central regions of the face, particularly the oral and nasal cavities, while the fibers in the dorsolateral tegmental sites were related to the cutaneous divisions mediated through the trigeminal tracts. The sensations which were referred by the patient to the perioral and nasal regions of the head could be the result of stimulating fibers of the dorsal ascending secondary trigeminal tract. This tract originates from the chief sensory nucleus and after crossing ascends through the midbrain in close proximity to the aqueduct. The “onion skin” pattern of sensory loss on the face that has been reported with certain lesions of the brain stem, such as a bilateral softening in the periaqueductal region, may be due to interference with this system of sensory fibers. Tactile sensations alone are thought to be mediated via this central route. Our patient described complex sensations of an unpleasant diffuse quality centered in and around the nose and mouth.

It was not always possible to determine on the basis of the subjective reports whether or not the lateral spinothalamic tract was activated alone or with the adjacent diffuse regions of the tegmentum. Differences were noted between the responses from the lateral tegmentum (spinothalamic) and the ventral tegmentum (lemniscal). At the level of the superior colliculus, the lateral spinothalamic tract appears on cross section as a small discrete bundle of fibers whose cross-sectional diameter has been greatly reduced as it ascends from the level of the cervical cord to the midbrain. This reduction in its size is presumed to be due to the occurrence of multiple synaptic interconnections throughout the brain stem.¹ The lateral spinothalamic tract is situated approximately 6 to 8 mm lateral from the aqueduct beneath the collicular plate. When the stimulations were confined to this region, the sensations evoked were always referred to wider zones on the body such as the opposite face and arm, or face, arm, chest, but rarely the leg. Our findings from stimulation suggest that this compact bundle of fibers considered to contain the major anatomical components of the spinothalamic tract must represent only a small fraction of the total ascending afferent systems which mediate painful sensations from the periphery. Both anatomical and neurophysiological observations support the presence of other diffuse ascending sensory systems passing through the mesencephalic tegmentum to the hypothalamus, thalamus, and cortex.^1,6,16 It is possible these dual systems of both the direct and diffuse afferent fibers remain in close proximity to each other as they ascend through this region of the mesencephalon.

In man, the feelings evoked from the central gray stimulation had a quality which he described as “fearful,” “frightful,” or “terrible,” and he would become apprehensive and not allow further stimulations. Our observations suggest that fiber systems in or near the central gray region have a topographic pattern of organization related to specific patterns of sensibility originating from the central regions of the body. The occurrence of both sensory and autonomic responses suggest a mixture of both afferent and efferent systems traversing this region. This central gray area appears to be important for the integration of complex sensory and motor activities related to feeding and sucking. The occurrence of the autonomic phenomenon elicited from these central regions may also indicate activation of the efferent fibers of the dorsal longitudinal fasciculus. This hypothalamic fasciculus represents a neural system for both efferent and afferent pathways connecting the hypothalamus with the caudal levels of the CNS. The anatomical location of fibers carrying gustatory and visceral sensations also have been localized to this periaqueductal region. They appear to have been activated by stimulation.

The evoking of at least two different kinds of sensation such as painful feelings and the unpleasant, fearful, and diffuse sensations from lateral and medial sites in the midbrain supports the idea that the transmission of painful or noxious impulses may take place through multiple pathways that are directed towards the midbrain, thalamus, hypothalamus, and the cerebral cortex.^1,6,13,14 Walker has already suggested that the midbrain may be one site for the conscious perception of the painful feelings in the thalamic syndrome.¹⁷ This idea was strengthened by our finding of several sites in the midbrain from which painful feelings were evoked, plus the simultaneous occurrence of certain types of pain with epileptic activity from focal regions of the mesencephalon. The hypothalamus may also be another site for the conscious perception of noxious stimuli since specific sensations were also evoked in man from the mesencephalic region through which the dorsal longitudinal fasciculus passes.^13,14,16

Changes in the physical and temporal patterns of electrical stimulation exert varied effects in the CNS and in its resultant behavior. Certain physiological effects were profoundly altered by changing the frequency of the stimuli. This has been true both in man and animals when the diencephalon and mesencephalon were activated.⁷ Delgado made a detailed study of these effects in the hypothalamus of the alert monkey.³ Nashold and Gills also noted that, following stimulations of the mesencephalon and diencephalon, there occurred a complex series of ocular, head, and neck movements which were observed with the higher frequencies (60, 120, 300 cps) of stimulation.⁸ Some surgeons using similar stereotaxic techniques have not consistently elicited pain during acute stimulation of the midbrain and few have elicited pain by stimulating the sensory thalamus.^4,9,11 This may be due in part to the use of low frequencies of stimulation which in our experience did not readily elicit pain. While we noted consistent complaints of pain from the mesencephalic stimulation with frequencies of 60, 120, and 300 cps, pain was not produced by thalamic activation under similar circumstances. The evocation of painful sensations may require a degree of summation of sensory input related to these higher frequencies.

Stimulation of the brain in man has had only limited application in stereotaxic operations. Complete information cannot be gathered at the time of the surgical operation where the comfort and safety of the patient is important.

The placement of therapeutic lesions in unexplored regions of the brain should be preceded by a careful analysis of its function by the use of chronically implanted electrodes. The techniques are safe. The accurate stereotaxic localization in a target area required a detailed knowledge of its regional neuroanatomy and its physiological function. A careful and systematic evaluation of these factors should be the goal of the stereotaxic surgeon.

The dorsolateral and ventral tegmental regions of the mesencephalon have been stimulated in man using chronic electrodes. The kind of sensation evoked was related to the locus of stimulation as well as the parameters of the stimuli. Activation of the dorsolateral tegmentum resulted in contralateral paresthesia or pain depending on the kind of stimuli. Central gray stimulations resulted in autonomic, emotional, and sensory responses which were referred to the midline regions of the body.