Cortical Mapping: Historical Perspective

Dr. Robert Bartholow performed the first electrical stimulation of the cortex in 1874 on a patient named Mary Rafferty, a 30-year-old Irish woman who had been employed as a domestic servant. She presented with an infected scalp ulcer, which was diagnosed as cancerous. The physicians attempted to treat this surgically, leaving a 2-in.-diameter hole in her skull with exposed dura. Apparently after determining that nothing could be done to save her life, Dr. Bartholow proceeded to experiment on the exposed brain, reportedly with the patient’s consent. By inserting needle electrodes into the exposed brain tissue and by applying small electrical currents to various areas, he noticed it caused movements in various parts of her body and did not cause pain, following the patient’s initial complaints of neck pain with needle insertion. He also noted that application of larger currents resulted in seizure activity and what seems to be what we now recognize as a transient postictal state. Although it is reported that she returned to consciousness 20 min later, she complained of weakness and vertigo. As her condition worsened, her physicians did not do any further experiments, and she died a few days later. The conclusions from her autopsy were that although parts of her brain had been damaged from the electrodes, her death was due to her cancer and not to this experiment. Despite this “contribution to medical science,” both British and American physicians severely criticized Dr. Bartholow and his “reckless use of living human beings,” and the American Medical Association condemned his experiments, calling them “so in conflict with the spirit of our profession, and opposed to our feelings of humanity that we cannot allow them to pass unnoticed” [1].

Despite this criticism, and the fact that this hardly could be considered “cortical mapping” in any sense of our current use of the term, this provided to beginnings of understanding that electricity could be applied to different areas of the brain and that regional somatic activity would result. Research on electrical stimulation of human and animal brains continued, and in 1888 Dr. Nancrede mapped the motor cortex by the use of a battery-operated bipolar stimulator probe. Neurologists David Ferrier and Victor Horsley used cortical stimulation mapping techniques to research the function of the precentral gyrus and the postcentral gyrus in the late 1800s. In the early 1900s Charles Sherrington used monopolar stimulation to elicit responses and was able to determine that the precentral gyrus elicited a motor response and that the postcentral gyrus was a sensory cortex. Dr. Harvey Cushing confirmed these findings and was the individual primarily responsible for moving cortical mapping from an experiment into an accepted neurosurgical technique.

Prior to going into the techniques used in the process of cortical mapping and analysis of those determinations, we should first have some understanding of what a cortical map is. Most physicians and medical students have seen the diagram of the homunculus, as described in the 1930s by a Canadian neurosurgeon, Dr. Wilder Penfield, and probably remember an image of a human body displayed across a drawing of a human cortex, although in a somewhat disjointed manner, with a large elongated face in the lateral one-third, while the hand is in the next approximately one-third, and the rest of the body is in the next one-third, toward the most central part, then with the foot on the medial portion of each hemisphere (see Chap. 2). Most physicians will probably also remember that the sensory functions are shown as existing in the postcentral gyrus of the parietal lobe, just posterior to the central (or Rolandic) sulcus, while the motor function in the cortex is described as primarily in the precentral gyrus of the anterior lobe, just anterior to the central (Rolandic) sulcus (with a very similar homunculus image). In effect, this displays the most basic concept of a cortical map. Unfortunately, however, this homunculus diagram is a grossly inaccurate oversimplification. The anatomical view of the brain tissue does not always precisely correlate to localizing the functions as suggested by this diagram. Sensory and motor areas of the cortex can now be mapped much more precisely by electrical stimulation and recording of “evoked” responses.

Anatomic and Physiologic Basis

The cortex of the human brain is 2–4 mm thick and in most parts of the cerebrum contains six layers which can rather easily be demonstrated on microscopic examination. In general, sensory cortex is thinner and motor cortex is thicker [2]. Within the brain cortex, very small areas (minicolumns) can be identified that perform a specific information processing function. Minicolumns grow from progenitor cells within the embryo and contain neurons within multiple layers (2–6) of the cortex [3]. A cortical minicolumn is a vertical column through the cortical layers of the brain, comprising approximately 80–120 neurons, except in the primate primary visual cortex where there are typically more than double this number. There are about 200,000,000 minicolumns in the human cortex. Many sources support the existence of minicolumns, especially Mountcastle [4] with strong evidence reviewed by Buxhoeveden and Casanova [5] who conclude “…the minicolumn must be considered a strong model for cortical organization” and that the minicolumn is “the most basic and consistent template by which the neocortex organizes its neurones, pathways, and intrinsic circuits.” It appears that this minicolumn structure is the primary means of organization in the cerebral cortex not only of humans but of other animals as well.

From multiple examinations and calculations, various researchers have estimated the diameter of a human minicolumn is about 28–60 μm. These minicolumns also contain downward projecting axons that are approximately 10 μm in diameter, with periodicity and density similar to those within the cortex, but not necessarily coincident. The probable estimated size of a minicolumn can also be calculated by area considerations: if the surface area of a human cortex (both hemispheres) is 1.27 × 10¹¹ μm² and if there are 2 × 10⁸ minicolumns in the cortex, then the cortical surface area of each minicolumn is 635 μm², giving an average diameter of 28 μm (but even if the cortex area were doubled to the commonly quoted value of 2.5 × 10¹¹ μm², this would rise to 40 μm). Johansson and Lansner do a similar calculation and arrive at an estimated minicolumn size of 36 μm.

There is also evidence from studies published in 2000 by two separate researchers, Buxhoeveden and Buldyrev, that spacing of 50–80 μm exists between adjacent columns. All cells in a single minicolumn have the same receptive field; adjacent minicolumns may have very different fields. Thus, a stimulus applied to a specific sensory nerve elicits a response within specific cortical minicolumns and does not necessarily elicit responses in immediately adjacent minicolumns in the cortex. This columnar arrangement forms the anatomic basis for the ability to perform cortical mapping. However, electrodes which are used for cortical mapping currently have a diameter of 2–3 mm, so we cannot electrically stimulate each discrete minicolumn but instead electrically stimulate a field containing hundreds of minicolumns with (hopefully) common functionality.

Maps of these cortical areas may be demonstrated in different ways such as texture maps, color maps, and contour maps. However, despite the existence of these maps, even for those attempting to localize the fields of “minicolumns,” which subserve a particular function in a human brain, it can be challenging. The brain retains a great degree of plasticity, such that if one of these areas is damaged, much of the function designated to that specific area can be “taken up” or assumed by a nearby area. Thus, designated maps can change with experience.

As an example of this plasticity phenomenon, people who read Braille (which is done with an index finger) develop large areas responsive to stimulation from the index finger. A homunculus mapped on the motor cortex of such a person would have a relatively huge index finger. This phenomenon contributes to the lack of accuracy and specificity of a “brain map” that the standard homunculus diagram would otherwise suggest is present in the human brain.

The cytoarchitecture of the cerebral cortex enables the recording of local positive and negative potentials over the cortical surface corresponding to the projection of cortical axons. As discussed in Chap. 4, this phenomenon is known as a dipole. The projection of dipoles varies among locations of the cortex, but the projection of the dipole of neurons in the primary somatosensory cortex (postcentral gyrus) is in the anterior–posterior plane. Furthermore the zero-potential or mid-dipole point lies over the central sulcus. As such, recording the cortical peak of the median nerve SSEP from a row of electrodes placed directly on the cortical surface is used as a means of locating the central sulcus and therefore both the primary motor and sensory cortical areas. The point at which a phase reversal (between positive and negative potentials) is seen can be reliably marked as the central sulcus. See Fig. 4.​3 in Chap. 4 for an example. Once the relative location of the motor cortex is identified, it can be further mapped as described below.

Equipment and Technique

So how is cortical mapping accomplished? The mapping is done during a craniotomy by stimulating the sensory or motor cortex with a weak electric current, usually for a few seconds, once the dura mater has been peeled back. This electrical stimulation acts as a transient reversible virtual lesion, interrupting the normal electrical activity in that localized area of neural tissue. This “lesion” can either induce or prevent a specific motor or sensory response that can be tested and evaluated. For example, the stimulation may produce tingling in part of the body, or movement in part of the body, or it can interfere with a normally spoken word.

As mentioned earlier, the electrodes currently used are usually circular with diameters of 2–3 mm. They are usually made of stainless steel or a platinum/iridium alloy and imbedded in a Silastic material. Due to difficulty in re-sterilizing them, they are single-use devices.

Stimulation for mapping is commonly performed according to one of two techniques. Using continual electrical stimulation (Penfield’s method), constant current is applied using a bipolar stimulating electrode at a frequency of 50–60 Hz. A biphasic square wave pulse with duration of 400–1000 μs is used in order to avoid charge buildup on the surface of the brain. A monophasic square wave pulse is not safe to use for this type of high-frequency continuous stimulation. A more modern stimulation technique used for mapping of the somatic motor areas is known as direct cortical electrical stimulation (DCES) or simply MEP mapping because of the similarity with transcranial motor evoked potential monitoring. DCES makes use of a pulse train as opposed to continuous stimulation. Monopolar anodal stimulation is used for DCES and due to the use of brief pulses; monophasic square waves are an acceptable stimulus. A train of 4–9 pulses with duration of 50–500 μs is usually effective. Electrodes may be placed individually or more usually in a row or in a grid array. The electrical current applied must be enough to stimulate the neurons for an adequate duration yet low enough to avoid damaging them. Whether using MEP mapping or Penfield’s technique , the “dose” of the current is usually started low and then gradually increased in both intensity and duration until a response is elicited. So initial intensities of 1 mA are a commonly used starting point. The current is then gradually increased by 0.5–1 mA with successive tests until a desired response is noted. It is important to identify the stimulation intensity that is adequate to produce activation of the neural tissue. Afterdischarges are nerve impulses that occur after stimulation, and the presence of afterdischarges indicates that the maximum amount of current that can be safely applied to the cortical surface has been reached. Monitoring for the presence of afterdischarges using electrocorticography (ECoG) is necessary to avoid the complication of seizure during cortical stimulation and also provides a measure of the adequacy of stimulation (see Chap. 10).

In situations where surgery needs to be performed to remove cerebral tissue, such as for tumor resection, or when an incision must be made through this more superficial cerebral tissue to get access to a deeper structure, a specific determination of the areas of the patient’s brain controlling a specific function becomes important. Likewise, it is also important to know where the “silent” areas are that surround these functions. The surgical goal may be to affect a particular cortical area or to specifically avoid affecting a few or many of these cortical areas.

Within this context, identification of “eloquent cortex” becomes quite important. Eloquent cortex is a term used by neurologists and neurosurgeons for areas of cortex that result in a loss of sensory processing or linguistic ability or some degree of loss of sensory or motor function if it is damaged or removed. These defined areas of cortex are crucial for certain particular functions, and some areas are indispensable for a particular cortical function [6]. The most commonly recognized areas of eloquent cortex are in the left temporal and frontal lobes, i.e., Broca’s and Wernicke’s areas (speech and language), bilateral occipital lobes (vision), bilateral parietal lobes (sensation), and bilateral motor cortex (movement).

Cortical mapping may also be done to attempt to identify an epileptogenic focus so that surgical excision or ablation of that area can be accomplished. The goal of complete resection of an epileptogenic focus must often be limited by sparing of eloquent cortex in order to avoid new and unacceptable deficits following epilepsy surgery. Although the homunculus diagram can provide a general idea of where specific motor or sensory functions are likely to be found in the cortex, intraoperative brain mapping provides much more specific information for a particular patient at that specific time of the surgical procedure. As suggested earlier, there are two broad areas of neurosurgery in which intraoperative cortical mapping is employed: excision of intracranial tumors and surgery to treat seizures.

Application in Cortical Tumor Excision

As already noted, anatomic appearance does not clearly and precisely identify areas of cortical brain tissue subserving a particular function. Multiple studies have shown that long-term prognosis is improved by more extensive tumor excision [7–10]. The use of intraoperative cortical mapping by electrical stimulation of specific anatomical areas provides the neurosurgeon with a real-time and patient-specific functional map. When using cortical mapping to identify eloquent cortex for tumor excision, the concept of positive mapping in contrast to negative mapping also comes into play.

A positive mapping occurs when eloquent areas are identified around the site of planned tumor excision. In other words, specific stimulation sites result in a recognized sensory or motor activity, and these sites are in close proximity to the area of resection of the neural tissue. A negative mapping occurs when electrical stimulation of a surrounding area does not produce any recognizable motor or sensory activity [11–13].

Although it would be easy to think that a positive identification of an area subserving a particular motor or sensory function would be desirable and would allow the surgeon to more precisely navigate the resection around it, experience has shown exactly the opposite. A negative mapping result around eloquent areas seems to provide a better “safe margin” for tumor resection with a low incidence of postoperative neurological defects [14]. In fact, positive identification of eloquent areas around the planned site of tumor resection actually increased the risk of postoperative deficits, probably indicating close proximity of tumor to functional cortex.

Application in Epilepsy Surgery

ECoG is employed in epilepsy surgery in an attempt to identify and remove the “epileptic zone” of tissue. This “epileptic zone ” is felt to be the anatomical site of seizure onset as well as the surrounding tissue which might potentially be recruited into the critical mass of tissue involved in the seizure. Although this technique essentially records the same type of electrical activity as an EEG, the electrode montage being placed directly onto the brain tissue, there is less attenuation and dispersion of the electrical signals. This is felt to provide more precise localization of the aberrant electrical activity causing the patient’s seizures than a diagnostic EEG.

ECoG requires the presence of a neurophysiologist interpreting the data in real time. Unlike other aspects of intraoperative monitoring, this cannot be accomplished through remote monitoring or telemedicine. The neurophysiologist must remain in close communication with the surgeon. A standard 16-channel EEG machine can be used to do the recording, but since the electrodes are directly on the brain, modifications from the normally used EEG settings of the recording sensitivities, time constants, filters, etc. are made. The machine is usually present in the operating room itself or in an operating room gallery with a two-way communication system in place so that the surgeon and neurophysiologist can communicate.

In order to accurately locate the epileptogenic area, the recording electrodes should be placed at equal distances from each other, both horizontally and vertically on the cortical surface. Angulated electrode placement should be avoided, since this can lead to false localization. Montages should contain at least four electrodes in a straight line.

Actual ictal events are rarely recorded. Instead, usually only interictal epileptiform activity is noted intraoperatively. The area of maximum epileptiform activity is felt to be the irritative zone that initiates the seizure, but this is not necessarily the area of origin of the epileptic seizure. Alacorn et al. found that removal of this area of maximal epileptiform activity yielded a better chance of good surgical outcome with reduction in epileptic activity, but if the area of maximal discharging was not completely resected, surgical outcome was likely to be poor.

Electrical stimulation of a suspected cortical area has been attempted but without good results in localizing the area of epileptogenic activity. When the stimulated area correlates with eliciting the typical aura preceding the seizure and this coincides with the area of greatest epileptic discharge, there seems to be a better correlation with successful surgical treatment of the seizure activity when that area is resected. However, if afterdischarges occur, the correlation to the epileptic zone is not as strong. This is probably due to afterdischarges originating from a distant and uninvolved area.

After surgical resection, sometimes residual spiking activity will occur. Unfortunately, the significance of this is not clear. While 75% of patients who were not seizure-free following resection had residual spikes noted on electrocorticography, 36% of patients who were seizure-free following surgical resection also exhibited residual spikes [15].

General Anesthesia

For sensorimotor mapping with the patient under general anesthesia, a total intravenous anesthesia regimen using propofol and narcotics is preferred. Neuromuscular blockade should be avoided. This regimen preserves the specificity of motor mapping while reducing the incidence of seizures in response to cortical stimulation. Although the use of propofol reduces the incidence of seizures, it does not eliminate the risk, and the team should be prepared to treat an intraoperative seizure if it occurs. The placement of bilateral soft bite blocks (as would be done for MEP monitoring) is also important.

Awake Craniotomy

A successful intraoperative course for an awake craniotomy starts with the preoperative evaluation. Medications should be noted, as well as concurrent medical conditions and serum levels of any antiseizure drugs currently being taken. A history of complications from medical management of seizures should also be discussed.

The patient should be given a detailed account of expected intraoperative events and warned of certain intraoperative events, such as opening the dura that may cause some discomfort. The anesthesiologist must reiterate the advantages of the patient being awake and ensure the patient that he/she will be present throughout the operation minimizing anxiety and pain when possible. Therefore, constant intraoperative communication with the neuroanesthesiologist will be expected. Lastly, induction and emergence from anesthesiology should be discussed with the patient.

Intraoperative: Local Anesthetics

Intraoperatively, the anesthesiologist and neurosurgeon use local anesthesia to perform regional, field, and dural blocks. Cutaneous nerves branching from the trigeminal nerve innervate the skin, scalp, pericranium, and periosteum. Subcutaneous infiltration with lidocaine or bupivacaine with epinephrine is commonly employed and successful in blocking afferent input to these areas. The skull has no sensory innervation, so it can be drilled and opened with no patient discomfort. The dura receives innervation from all three divisions of the trigeminal nerve, the recurrent meningeal branch of the vagus, and by branches of the upper cervical roots and can produce significant discomfort for the patient when instrumented by the surgeon. Local application by the surgeon can work; however, if this becomes too unpleasant for the patient, then general anesthesia can be induced and a laryngeal mask airway inserted until exposure is completed.

Intraoperative Sedation

Current techniques commonly use propofol, fentanyl, remifentanil, or dexmedetomidate. Many use propofol infusions with slow and careful titrations of fentanyl. Most recently remifentanil has replaced fentanyl due to its ultrashort action and is combined with propofol to provide sedation and analgesia during awake craniotomies. This technique is popular because of the safety profile and lack of respiratory depression if carefully titrated. Dexmedetomidate , an alpha-2-adrenoreceptor agonist , has gained popularity due to its ability to provide analgesia and sedation, which is easily reversed with oral communication. Additionally, it produces no respiratory depression when used alone. Ensuring sedation and analgesia for the patient while preventing apnea or airway obstruction is the main concern for the anesthesiologist. Airway equipment (oral and nasal airways, laryngeal mask airways, and emergency intubation equipment) must be readily available throughout the case.

Asleep–Awake–Asleep Technique

Propofol and remifentanil are the two agents used most frequently for this technique. Both are short-acting, safe, and predictable. Additionally, they can be titrated while using the bispectral index monitoring system, which provides the anesthesiologist more precision for drug dosing adjustment. A laryngeal mask airway is commonly inserted to prevent airway obstruction. With proper propofol and remifentanil dosing, airway irritation is alleviated and neuromuscular blocking agents are not warranted. After the craniotomy is completed and the dura is opened, the remifentanil dose is reduced or stopped, and spontaneous respirations are allowed to resume. The propofol infusion is reduced, and the LMA is removed as the patient regains consciousness. After surgical resection is completed, the infusions are reinstated and the LMA placed until surgery is completed.

Contraindication to Awake Craniotomies

Multiple issues must be considered before proceeding with an awake craniotomy. The ability to communicate with the surgeon and anesthesiologists is imperative. Any communication problems, such as dysphasia, are strong contraindications for awake craniotomies. Extremely anxious patients or patients prone to an exaggerated pain response should probably be avoided as are patients requiring prone positioning for surgery. Patients with lesions requiring extensive dural surgical resection should probably be avoided. Finally, lengthy surgical procedures may make it difficult for patients who are required to lie still.