A Biocompatible Titanium Headpost for Stabilizing Behaving Monkeys

Daniel L. Adams, John R. Economides, Cristina M. Jocson, Jonathan C. Horton


Many neurophysiological experiments involving monkeys require that the head be stabilized while the animal performs a task. Often a post is attached to the skull to accomplish this goal, using a headcap formed from dental acrylic. We describe a new headpost, developed by refinement of several prototypes, and supply an AutoCAD file to aid in machine shop production. This headpost is fabricated from a single piece of commercially pure titanium. It has a footplate consisting of four limbs arranged in the configuration of a “K.” These are bent during surgery to match the curvature of the skull and attached with specialized titanium bone screws. Headposts were implanted in seven rhesus monkeys ranging in age from 2 yr to adult. None has been rejected after up to 17 mo of regular use. They require little or no daily toilette and create only a 0.80-cm2 defect in the scalp. Computed tomography after implantation showed that the skull undergoes remodeling to embed the footplate in bone. This finding was confirmed by necropsy in two subjects. The outer table of the skull had grown over the titanium footplate, whereas the inner table had thickened to bury the tips of the titanium screws. The remarkable strength of the skull/implant bond was demonstrated by applying increasing amounts of torque to the headpost. At 26.3 Nm, the headpost tore from its metal footplate, but no screws came loose. The excellent performance of this implant is explained by integration of biocompatible titanium into remodeled bone tissue. The headpost is simpler to implant, more securely anchored, easier to maintain, and less obtrusive than devices attached with acrylic.


Neural recordings in “awake behaving” or alert monkeys have driven huge advances in neuroscience research. The technique allows correlation of the electrical activity of single cells with perception and behavior. Currently, no other technique matches both the spatial and temporal precision of physiological recording using extracellular electrodes.

To perform such recordings two devices are often implanted: a headpost and a recording chamber. The headpost is useful to keep the head stationary because sudden rotations cause the brain to move relative to the electrode tip, destabilizing the recording. Even if electrode recordings are not being made, a headpost is useful because it makes monitoring of gaze position during psychophysical experiments much easier by eliminating head movement. The chamber serves two purposes: it provides resealable access to the brain (usually exposing only the dura) and a stable structure onto which the electrode microdrive can be mounted. Usually, the headpost and the recording chamber are embedded in a headcap. Dental acrylic is commonly used to form the headcap, but it does not adhere to the skull sufficiently to be used alone. Metal screws or bolts must be placed in the skull to form a framework to anchor the acrylic. This basic technique has been refined further by using bolts designed to minimize the pressure and stress exerted on the bone and by incorporating a metal headpiece that distributes the forces evenly over the implant (Betelak et al. 2001; Lemon and Prochazka 1984; Porter et al. 1971).

The acrylic-based implant is adequate and widely used. However, it is far from ideal. The most problematic component is the dental acrylic itself. Most commercially available dental acrylics are formulated from methyl methacrylate that polymerizes in the presence of an initiator (benzoyl peroxide). The polymerization reaction is strongly exothermic so care must be taken to avoid increases in temperature during application that can damage the bone. There is evidence that the monomer form of acrylic cement, which can leak from the headcap, is toxic to bone (Albrektsson and Linder 1984). The finished acrylic cap may cover much of the animal's head. Thus a large surface area of bone must be stripped of soft tissue. Sometimes, bonding of the acrylic to the skull is encouraged by scoring the bone surface to provide a “key.” These procedures damage the bone and then prevent natural healing by forming an acrylic barrier between the bone and the periosteal and connective tissues. The large cap also results in a long shoreline between the scalp incision and the implant. This is a prime site for bacterial infection and may require daily cleaning with antiseptic agents and treatment with antibiotics. Because acrylic bonds poorly to healthy bone, a capillary gap between the cap and the skull surface can arise, providing an ideal environment for bacterial colonization. Once infected, the bone becomes necrotic, resulting eventually in expulsion of the retaining screws and failure of the implant.

To overcome these problems, a number of laboratories have developed alternative methods of head fixation (Betelak et al. 2001; Evarts 1968; Foeller and Tychsen 2002; Isoda et al. 2005; Lemon and Prochazka 1984; Pigarev et al. 1997; Yakushin et al. 2000). Potentially, the elimination of acrylic could result in shorter, easier implant surgery and a reduction in overall costs. Furthermore, without a large expanse of acrylic to drill through, the removal of a recording chamber and the implantation of further chambers onto the same animal would be greatly simplified. A more biocompatible headpost device would also improve the quality of life for monkeys used in neuroscience research. The ideal acrylic-free implant is comfortable, stable, durable, sanitary, unobtrusive, and requires no maintenance. Here, we share our experience with the design and implantation of acrylic-free titanium headposts in the hope that other investigators will be encouraged to adopt and further develop this important refinement.


Headpost design

The objective was to provide a robust and stable method of head fixation in macaques without the use of dental acrylic. This was achieved by using a headpost with a footplate that was bent intraoperatively to match the curvature of the skull. It was fastened to the skull with surgical cortex screws. Titanium was used, primarily because it has the remarkable property of being readily osseointegrated into healthy bone tissue. Osseointegration is the process whereby direct structural connections are made between the living bone and the titanium surface (Brånemark et al. 1969; Linder et al. 1983). Titanium is commonly alloyed with aluminum and vanadium (e.g., 90% Ti, 6% Al, 4% V), resulting in a near doubling of its tensile strength. However, titanium alloys are less ductile and harder to machine than pure titanium (www.titaniuminfogroup.co.uk). For this reason we used unalloyed, commercially pure (CP, grade 2) titanium for headpost manufacture. It has a low density (4.5g cm−3, 60% that of steel) and high tensile strength (345 MPa, a value comparable to that of standard structural steel). Although not as strong as most titanium alloys, CP titanium has excellent osseointegrative properties and is ideal for use as a load-bearing implant (Pohler 2000; Rubo de Rezende and Johansson 1993).

Our experience with titanium headpost design began with a device obtained commercially (Part #6-FHP-X2F, Crist Instrument, Hagerstown, MD). The headpost had an “X”-shaped footplate designed for attachment to the skull with 12 bone screws. The vertical post had a tapered cross section, designed to mate with a headpost holder (Part #6-FHB-S2B, Crist Instrument). The post and its footplate were manufactured separately and welded together. Titanium welding is a difficult process because it is prone to contamination, which renders the weld brittle. This headpost was implanted in a single adult male rhesus weighing 8 kg. Within a short time the titanium weld failed, suggesting that it would be preferable to machine the post and footplate from a single piece of metal.

The first design (the Mark I model) specified that the headpost be turned from 2.5-in.-diameter CP titanium bar stock. The shape of the footplate was changed from an “X” to a “K” shape, allowing the headpost to be positioned closer to the orbital ridge, with the two co-linear feet placed anteriorly. The number of screw holes was increased from 12 to 14, by adding one to each of the oblique limbs. Headposts with this design were implanted into four male rhesus monkeys: two adults (weight: 10 and 9 kg) and two juveniles (weight: 4.0 and 6.5 kg).

Following these successful implantations further improvements were made, resulting in the Mark II headpost design. The post was reduced in length and the feet were made longer. Four screw holes were placed in each limb, increasing the total number from 14 to 16. The distance between the screw holes was increased slightly to reduce the possibility of weakening the bone by “chain-drilling.” The pointed end of the post and a stepped shoulder at the interface between the post and the footplate were both rounded out. The footplate screw holes were also countersunk more deeply to lower the profiles of the screw heads. The Mark II design is shown in Fig. 1. It is available from on-line supplementary material as an AutoCAD document (Autodesk, San Rafael, CA). This headpost design was posted at www.mfg.com, a website for sourcing projects to machinists, and sent to a computer numerical control (CNC) workshop for production. It has been implanted into one adult male rhesus monkey weighing 10 kg and one adult female rhesus monkey weighing 6 kg.

FIG. 1.

Third angle orthographic projection of the Mark II headpost. It is machined on a computer numerical control (CNC) mill from a single piece of commercially pure (CP) titanium. Footplate is 1 mm thick and has 4 feet arranged in a “K” shape. Each foot has 4 screw holes, countersunk to accept orthopedic titanium screws. Post has a 14° taper angle that is cut using CNC–electrical discharge machining (EDM). It mates with a similarly tapered slot in the Crist Instrument headpost holder. Larger of the 2 holes in the post accepts an indexing pin to align the smaller tapped hole with a hole in the headpost holder. A 6–32 screw passes through this hole and into the post. When tightened, this screw locks the post to the headpost holder.

The Mark I manufacturing process was expensive because a large amount of material had to be removed from the bar stock to form the post. To reduce the machining time required for roughing out, the Mark II was manufactured from a 2-in.-thick titanium plate on a milling machine, rather than from round bar on a lathe. The tapered faces of the post were cut using electrical discharge machining (CNC-EDM). The final product was grit-blasted (mesh size 50) to roughen the surface finish and encourage osseointegration (Buser et al. 1991). The total weight of the Mark II headpost was 9.4 g (not including the 16 titanium screws).

Implantation surgery

All procedures were approved by the UCSF Institutional Animal Care and Use Committee. The animal was sedated with ketamine [10 mg/kg, administered intramuscularly (im)] and prepared for sterile surgery. An endotracheal tube was inserted and the animal was ventilated with 2% isoflurane in 50/50 N2O/O2 to maintain surgical anesthesia. EKG, rectal temperature, inspired and expired anesthetic agents, end-tidal CO2, and oxygen saturation were continuously monitored. The animal was placed on a heated blanket in a stereotaxic instrument. The scalp was shaved and swabbed with 5% povidone-iodine. A sagittal incision was made, starting approximately 5 mm behind the brow ridge and extending about 50 mm posteriorly. Bare bone was exposed over a region measuring about 60 × 30 mm by scraping back the periosteum. In some animals, particularly adult males, it was necessary to disinsert each temporalis muscle from the superior temporalis line to create enough space for the footplate. This was done with a small bone curette to avoid damage to the muscle. The muscle spontaneously reinserted over the footplate after surgery.

The headpost was offered up at a location about 5–10 mm posterior to the brow ridge and the feet were bent to match the curvature of the skull (Fig. 2). To encourage the feet to curve evenly along their length, rather than at a single point proximal to the post, they were struck gently with a hammer against a 35-mm-diameter round steel anvil. We also used a metal bar with a 1.5-mm slot in the side to bend the plate during surgery. The headpost was immersed in a sterilizing solution before placing it back on the skull to test the fit. Once the feet were suitably curved to conform to the skull surface (maximum ∼2-mm gap), the headpost was held in its final position by hand.

FIG. 2.

Surgical photograph showing the midline incision held open with retractors, exposing the skull surface. Crist Instrument headpost is shown bent to match the profile of the skull and held in place with one screw. Dashed line marks the midline of the skull. Brow ridge is visible toward the bottom right of the image.

The centers of two proximal screw holes on opposite sides of the post were marked on the skull using a surgical pen. It was crucial to accurately position the first two pilot holes to avoid a crooked headpost. A 2-mm orthopedic drill bit (Part #2.0 mm QCK, Veterinary Orthopedics Implants, South Burlington, VT) was used to make pilot holes for the screws. The holes were drilled down to the inner table of the bone, taking care not to damage the underlying dura. It was important not to subject the bone to excessive friction during drilling because heat can cause softening of the bone surrounding the hole, resulting in eventual screw ejection. Heating of bone by only 7–8° above normal body temperature for 1 min has been shown to kill osteocytes and to retard healing (Eriksson and Albrektsson 1983). For this reason, we held the drill bit in a “T”-type tap wrench and turned it at about 50 rpm by hand while applying light pressure and irrigating the hole with sterile saline. It was possible to judge exactly when the pilot hole had the correct depth because the instant the drill bit's point perforated the inner table a sudden increase in turning resistance was felt as a result of the drill bit biting deeper into the bone. The correct depth was verified by observing through an operating microscope a slight fluid pulsation through the small communication with the epidural space. Alternatively, we gently probed with a blunt dental pick to detect a soft spot at the center of the hole. The full diameter of the drill bit should not traverse the inner table of the skull. Once an estimate of the correct pilot hole depth was attained, we attached a collar to the drill bit as a safety measure to guard against slippage into the brain.

At this point, the headpost was held in place and the first two screws were started in their pilot holes. We used 2.7-mm-diameter × 8-mm-long titanium cortex screws (Part #T270.08, Veterinary Orthopedics Implants). These screws have a dome-shaped tip that depresses, rather than penetrates, the dura mater (Fig. 3). The screws were driven using a 2.5-mm Allen key inserted into a watchmaker's screwdriver handle. The key to successful screw insertion was to apply ample pressure and to turn slowly, ensuring that the screw advanced further into the pilot hole with every small rotation. Overtightening will strip the thread made by the screw, evinced by the screw rotating freely in the hole without advancing further. To prevent this problem, we avoided applying torque once the screw head was in contact with the headpost footplate. To ensure that the headpost remained in its intended location, the first two screws were tightened in tandem. The remaining holes were drilled in a proximal to distal order, each centered within a footplate hole. A screw was driven into each pilot hole before the next hole was drilled to minimize any cumulative misalignment as the footplate formed itself better to the contour of the skull. To gain access to the most distal footplate holes while keeping the scalp incision to a minimal size, it was useful to retract the scalp margin with a lever pivoted against the end of the footplate. During the drilling and screwing phase, we periodically checked the tightness of screws already in place, taking up any slack caused by the setting down of the footplate limbs. It was not necessary to manually tap the holes or to use self-tapping screws. This also avoided bringing the sharp edge of a tap or of a self-tapping screw into contact with the dura. Finally, when all screws were in place, we made several rounds of incremental tightening, taking care not to overtighten any screw. The maximum torque we applied was about 0.4 Nm.

FIG. 3.

Schematic drawing to show bone drilling and screw insertion. Pilot hole (2 mm) is drilled by manually turning the drill bit in a tap wrench. Hole passes through the outer table of the skull and perforates the inner table (left). On insertion of the screw (middle), remaining bone fragments are pushed through, to become incorporated into deposits of woven bone during the healing process. After healing (right), the screw becomes encapsulated completely in lamellar bone tissue.

After ample irrigation with physiological saline, the galea was sutured over the top of the footplate using 5–0 or 6–0 Polyglactin (Vicryl, Ethicon, Somerville, NJ) absorbable suture. It was important to close the galea snuggly around the base of the post to avoid later skin retraction. The skin incision was closed with 4–0 silk, drawing the scalp up tightly to the base of the post. After cleaning the skin and applying a topical antibiotic ointment (neomycin sulfate, polymyxin sulfate, and bacitracin zinc, Vetro-biotic, PharmaDerm division of Altana, Melville, NY), the animal was recovered from anesthesia. Buprenorphine HCl, an opiate analgesic (Buprenex, 0.3 mg kg−1 im), was given every 8 h for ≥48 h, until the potential for postoperative pain had passed. The scalp incision healed quickly and external sutures were removed after 1 wk. We waited 2 wk for the headpost to stabilize before fixing the animal's head for the first time.

Use in juvenile animals

Some studies require head stabilization in young animals. Attachment of a headpost in a young monkey using screws cannot be done if the skull is too thin or poorly ossified. To address this issue, we prepared a dry skull specimen from a male rhesus monkey aged 26 mo. A paper template of the footplate was placed on the frontal bone at the location intended for headpost implantation. Then, the thickness of the skull was measured at the site of each planned screw hole by placing the depth gauge of a caliper through the foramen magnum. For comparison, bone thickness measurements were made using the same technique in dry skull specimens from two adult male rhesus monkeys and four adult female rhesus monkeys.

After we ascertained that the skull was thick enough to support a screw-mounted footplate, we implanted the Mark I titanium headpost in two juveniles, aged 29 and 38 mo. For the younger animal we used a pediatric version of the headpost, with a smaller footplate containing only 12 screw holes.

Assessment of performance

We photographed the headposts at regular intervals after implantation. Two animals reached the end of their experimental careers. They received a lethal dose of sodium pentobarbital (150 mg kg−1 administered intravenously), followed by perfusion with normal saline and 1% paraformaldehyde. Their brains were kept for histological analysis and an intact plate of frontal bone with the implant still attached was removed. The footplate was cleaned of soft tissue and photographed to assess the extent of bone growth. In some cases, the bone/screw interface was examined by cutting a cross section centered on a screw hole using a miniature circular saw.

The strength of the Mark I implant–bone interface was assessed by embedding the frontal bone in an epoxy resin block (Fig. 4). Contact between the epoxy and the titanium hardware was prevented by surrounding the tips of the screws and the “K”-shaped footplate with modeling clay. Once cured, the block was bolted to a vertical surface. It was oriented such that the post was horizontal and the medial–lateral skull axis was vertical. A 25-cm-long 2.5-cm diameter aluminum shaft (Part #6-FHB-S2A, Crist Instrument) was attached to the headpost by the headpost holder (Part #6-FHB-S2B, Crist Instrument). A 19-L plastic bottle was then suspended from the end of the aluminum bar at a distance of 29 cm from the headpost base. The bottle was gradually filled with water by a hose. The mass suspended from the aluminum bar at the moment of headpost failure was used to calculate the torque in Nm (newton-meter).

FIG. 4.

Method for testing the strength of the Mark I implant. A: schematic diagram of the macaque bone plate and attached headpost in the testing rig. Bone plate was embedded in epoxy resin to hold it securely during testing. Titanium headpost and screws were shielded from the epoxy resin with modeling clay to prevent adhesive reinforcement of the headpost attachment (see circular inset for detail). Epoxy block was bolted to a vertical wooden board. A cable was looped over the headpost holder bar at a distance (d) of 0.29 m from the post/footplate interface. A 19-L plastic bottle (from a drinking water dispenser) was suspended from the cable. Water was piped continuously into the bottle at a rate of about 2 L/min, creating a linear increase in the downward force (f) of about 20 N/min. Water supply was stopped at the moment of headpost failure and the bottle was weighed. Force required to break the headpost was calculated in terms of torque (d × f) measured in Nm, taking into account the headpost holder and shaft. B: photograph of the implant in the rig before testing. C: photograph of the implant after the test. When the torque reached 26.3 Nm (corresponding to a mass of 8.78 kg applied 0.29 m from the fulcrum), the 1 × 14-mm titanium plate fractured, causing the post to separate from the upper 2 footplate limbs. Screws and bone remained intact.


The commercial titanium headpost (Part #6-FHP-X2F, Crist Instrument) was implanted in a single 8-kg adult male rhesus. The animal had been chair-trained previously for 18 mo and was well accustomed to the experimental rig and to human contact. The headpost was used for the first time 12 days after surgery. Unfortunately, after 6 days of 2-h chairing sessions with head restraint, the headpost failed. The post broke off the footplate cleanly at the welded joint, leaving the “X”-shaped footplate attached securely to the skull (Fig. 5A). The animal was not harmed. However, it was deemed not prudent to remove the broken footplate and to implant another headpost. Thus the animal could no longer be used for awake behaving experiments. The scalp was sutured closed over the footplate. Ten days later, the monkey was used for a terminal physiology experiment under general anesthesia. At our suggestion, Crist Instrument Company no longer makes welded headposts. They are now fabricated from a single piece of metal.

FIG. 5.

Crist Instrument headpost footplate 1 mo after implantation surgery. A: dorsal view of the skull at necropsy with the “X”-shaped footplate still attached. Central hole in the footplate was produced when the weld between the post and the footplate failed, causing the post to fracture off. B: view from the underside of the skull plate, showing the protruding tips of 12 bone screws. C: dorsal view of the skull with the footplate removed. Deposits of woven bone are evident on the skull formerly covered by the footplate (arrow). Woven bone appears lighter in color and less smooth than normal skull. D: high-magnification view of boxed region in B showing the buildup of woven bone around the tips of the screws as they emerge inside the skull. E: cross section through the skull made at the location of the line in B and C. Bone that was in contact with the screw thread appears smooth, hard, and healthy. No signs of heat or pressure-induced necrosis are evident. F: same section shown in E. Screw has been replaced to demonstrate the tight fit between screw thread and bone. Dashed lines show the interface between the original bone surface and bone deposited in response to the implant.

After <1 mo with the Crist headpost in place, it was possible to see signs of bone healing and remodeling (Fig. 5). The inner and outer surfaces adjacent to the implant and screws were thickened with calluses of “woven bone”—a fast-forming calcified tissue produced by osteoblasts in response to injury. Woven bone was particularly evident adjacent to the screw tips, on the inside of the skull, where it had begun to build up a fillet around each protruding screw (Fig. 5D). It was also present on the outer surface of the skull, underneath the footplate, indicating that the close apposition of titanium alone is sufficient to stimulate callus formation. In normal bone healing, the woven-bone calluses are replaced by dense load-bearing lamellar bone over a period of a few weeks (Albrektsson 1992; McKoy et al. 2000).

A criticism that has been leveled at the use of bone screws for implantation is that a securely mounted screw exerts pressure on the surrounding bone, causing local necrosis, softening of the bone, and eventual ejection of the screw (Lemon and Prochazka 1984). To examine the screw/bone interface for signs of necrosis, we removed the screws and footplate and cut a section centered on a screw hole through the skull (Fig. 5E). On removal, we found that the screws were held firmly in place and unscrewed smoothly without damaging the bone. The inside surface of the hole appeared glossy and healthy. There were no signs of necrosis or fibrous tissue. The bone had sealed around the screw and formed a hard threaded surface. Woven-bone callus formation was observed on both the inside and outside of the skull in the vicinity of the screw (Fig. 5F). The same observations were made for all the other 11 screw holes.

After the two-piece welded titanium headpost failed, we implanted a single-piece headpost, the Mark I. Figure 6 shows the Mark I headpost in place on an adult macaque. The photograph was taken 4 mo after implantation surgery. By this time, the surgical incision had closed around the post and was no longer visible through the fur. The scalp covered the footplate entirely and was tightly apposed to the base of the post. There was no discharge, inflammation, or other evidence of infection. The animal showed no signs of discomfort during manual manipulation of the post or during daily experimental sessions with the head restrained. Except for the application of a topical antibiotic for 1 wk after surgery, no maintenance or cleaning was performed throughout the lifetime of the implant.

FIG. 6.

Mark I headpost 4 mo after implantation surgery. At this time, the animal was being chaired with its head held comfortably on a daily basis. Footplate was entirely covered by healthy scalp. Close proximity of bone and scalp anterior and posterior to the post allowed the scalp to heal close to the base of the post. This headpost required no maintenance.

The Mark I headpost shown in Fig. 6 was in place for a total of 14 mo. During this time the animal was chaired and the head fixed on a regular basis. No problems with the headpost were encountered during the animal's tenure in the laboratory. The monkey was perfused at age 9 yr and a plate of skull removed with the headpost still attached (Fig. 7). The dura mater showed no thickening or reaction from the proximity of the titanium screw tips. Examination of the implant specimen showed that there had been bone growth over the screws and footplate. The bone had grown around the implant feet, nearly bridging them completely in places (Fig. 7C). Many screw tips that originally protruded through the inner table were buried by newly formed lamellar bone. Other screws had only their domed tips visible (Fig. 7, B and D). There was also a marked thickening of the bone in regions adjacent to the footplate (Fig. 7, E and F). Comparison of skull thickness measured in the same locations in a male macaque of similar age showed that about 2–3 mm of lamellar bone tissue had been added to the outer table in the implanted animal.

FIG. 7.

Mark I headpost 14 mo after implantation surgery. A: dorsal view of the bone flap with headpost still attached. Areas where the bone has enveloped the footplate can be seen on all 4 limbs. B: view of the inside of the skull. Seven of the 14 screws are entirely covered by new bone growth. Only the tips of the remaining 7 screws are visible. C: high-power view of the boxed region in A, showing the extent to which the limb of the footplate has been assimilated into the skull. D: high-power view of the boxed region in B, showing how remodeled bone has enveloped the protruding screws. E and F: low- and high-power views of the anterior surface of the bone plate, showing extensive thickening of the outer table in the vicinity of the footplate. Remodeled bone is dense, load-bearing lamellar bone, easily distinguishable from the darker trabecular bone of the diploë.

We used the headpost and bone plate shown in Fig. 7 to test how much torque the implant could withstand. Torque was applied to the headpost by adding water to a plastic bottle suspended from the end of the headpost holder (Fig. 4). At a torque of 26.3 Nm (19.4 ft-lbs) the titanium footplate fractured, disconnecting it from the footplate on one side and bending the footplate on the other. For comparison, this is roughly equal to the recommended torque used to tighten an automotive spark plug. Inspection of the broken headpost showed that a crack had developed on the side where the footplate was being pulled away from the skull. The crack initiated at the sharp inside corner of the post–footplate interface and spread outward until the post was torn free from both limbs. None of the screws pulled out of the bone and the footplate remained firmly attached to the skull (Fig. 4C).

After the successful implantation of an adult macaque with the Mark I headpost, we decided to try the Mark I model in a juvenile animal. Concerned that the skull might be too thin to support the insertion of screws, we measured its thickness at the locations of the intended screw holes (see methods). The skull thickness ranged from 1.38 to 3.16 mm (mean 1.95 mm). For comparison, we made similar measurements in six intact dry skulls from adult rhesus monkeys. In two male specimens the frontal bone thickness at screw-hole sites ranged from 1.93 to 4.53 mm (mean 2.93 mm). In four females it was 1.04–3.41 mm (mean 2.23 mm). Thus the 2-yr-old macaque's skull was only about 1 mm thinner than the adults' skulls. Given this small difference, we thought it safe to implant two juvenile (29 and 38 mo old) macaques. For the younger of these two animals, we used a pediatric version of the Mark I implant, having shorter footplate limbs with three screws holes each. The implant was attached using mostly 8-mm screws; some 6-mm screws were also used, where the skull was thinnest. One 6-mm screw, located in a hole directly adjacent to the post, extruded spontaneously several months later. Based on this experience, it seems advisable to use 8-mm screws, even in young monkeys. This animal's headpost has functioned well during hundreds of daily experimental sessions, despite being held in place by only 11 screws. A 2- to 3-mm collar of exposed granulation tissue is present around the base of the post, which is cleaned weekly with a disinfectant consisting of a 1:1 mixture of 3% hydrogen peroxide and 5% povidone-iodine.

A computed tomography scan (Fig. 8) was obtained 15 mo after headpost implantation in this juvenile monkey, showing that the growth of the skull was not restricted by the footplate. Considerable remodeling of the bone occurred in response to the titanium hardware. In places, the footplate was covered with a layer of bone up to 1 mm thick. New bone deposits were also evident on the inside of the skull, surrounding the screw tips that originally protruded into the skull cavity. This skull remodeling was similar to what we observed in the postmortem specimen (Fig. 5).

FIG. 8.

Computed tomography images of the Mark I implant in place on a male macaque. Animal was implanted at age 29 mo and imaged at age 44 mo. A: lateral “X-ray” image of the whole skull showing the anterior–posterior location of the implant. Soft tissue is scarcely visible. Some bone thickening can be observed immediately in front of and behind the footplate. B: surface rendering of the skull and implant. Distal footplate limbs appear sunken into the skull surface. C: coronal 0.8-mm-thick section centered on the row of screws in the anterior footplate limbs. Soft tissue is dark gray, bone is light gray, and titanium is white. Six screws are visible; their hexagonal cups are seen in cross section. Considerable bone growth has occurred on both the inside and outside of the skull. In places, the footplate has been buried by ≤2 mm of newly formed bone. Tips of the 8-mm screws that originally protruded into the skull cavity have become embedded in bone. Scale bar applies to C and D. D: parasagittal slice (0.8 mm thick) taken through the right distal end of the implant footplate, angled to be axial to the screws. In this region, the bone remodeling in response to the implant was greatest, resulting in nearly complete coverage of the titanium footplate.

A headpost was implanted into another macaque at age 38 mo. The animal has been chaired and restrained on a daily basis for 9 mo since surgery. The animal shows no signs of discomfort and the implant is secure and clean.

Finally, two Mark II headposts (Fig. 1) have been implanted in adult macaques. These remain firmly attached 7 and 1 mo after placement and the animals are healthy and comfortable. Both animals are being used for daily experiments with head restraint. In the animal implanted 7 mo ago there is a 2-mm gap between the skin and the headpost, which is swabbed occasionally with a cleaning solution. This procedure is done without anesthesia while the headpost is attached to the chair and the monkey shows no signs of stress or discomfort. In the more recently implanted monkey, the skin is closely apposed to the base of the post.


We have implanted titanium headposts in seven macaques, using a simple method that attaches the footplate directly to the skull with bone screws. The device was placed in the midline, so that none of the screws came close to the superior sagittal sinus. After implantation, the presence of a titanium foreign body stimulated skull remodeling. The bone grew over the implant footplate and proliferated around the screws, gradually embedding the headpost in the skull. This resulted in a strong joint between the bone and metal. Titanium causes some artifact in magnetic resonance imaging or computed tomography, but less than that of other metals such as stainless steel.

We measured the amount of torque required to tear a headpost loose from a skull in a specimen obtained from a macaque 14 mo after implantation. After 26.3 Nm was applied the titanium headpost failed, but the screws did not pull out of the skull. This is testimony to the remarkable strength of the attachment to the skull provided by the healing of the screws into the bone. It is unlikely that a chaired animal with limited neck movement would be capable of generating a force equal to 26.3 Nm. However, caution is recommended when connecting the headpost to the post holder. In refining the headpost, we did not change the design of the mechanism for attaching it to the post holder. The Crist Instrument system uses a tapered headpost that mates with a slot in the headpost holder. The slot contains an indexing pin to align the post with a 6–32 screw, inserted through the post holder and into a tapped hole in the headpost (Fig. 1). The hole that accepts the indexing pin generates a weak point in the headpost. The cross-sectional area of the post at the center of the hole is reduced from 46 to 14 mm2. For this reason, it is important to attach the headpost holder (Part #6-FHB-S2B, Crist Instrument) to the headpost before connecting the heavy aluminum shaft (Part #6-FHB-S2A, Crist Instrument). Otherwise, there is some risk of bending or cracking the top of the headpost at the site of the hole for the indexing pin.

The tearing of the metal at the base of the Mark 1 headpost after application of torque showed that the implant was limited by the strength of the titanium, not the grip of the screws to the bone. The implant could have been strengthened by making the footplate limbs wider or thicker. However, this change might make it more difficult for bone to grow over the footplate or for scalp to close around the headpost shaft. Therefore the Mark II headpost was designed to strengthen the implant by creating a smooth radius at the interface between the headpost and the footplate to dissipate stress more effectively.

We hesitated to implant the titanium headpost in a 2-yr-old monkey, but the device has functioned without difficulty, despite the fact that a juvenile's skull is approximately 1 mm thinner than an adult's skull. It remains to be determined whether the headpost can be implanted safely in monkeys younger than age 2 yr. Note that one can easily convert the Mark II headpost to a pediatric size by filing off the distal screw holes from each limb of the footplate.

The use of acrylic-free implants for stabilization of the head in alert monkeys is not a new innovation. However, considering the advantages it offers over traditional methods, it has been slow to be adopted. There may be a number of reasons for this, but chief is a general unwillingness to experiment with an unproven technique, when a tried-and-tested one is available. This is particularly true in the case of primate research, where months of training can be invested before an experiment can commence. This factor, coupled with the high cost of animals and their care, makes it risky to experiment with changes in technique. Our aim in writing this report is to describe what we have learned through personal experience. We have also conducted a survey of other laboratories to share the experience of other scientists who have used similar footed titanium posts.

A list of 63 laboratories using awake behaving macaques in vision research in the United States was compiled by consulting PubMed. A brief survey was sent by E-mail to all 63 labs, followed by a reminder to those laboratories that did not respond. Replies were received eventually from 48 laboratories. A total of 137 footed, screw-mounted headposts have been implanted by 12 laboratories. Of 30 welded headposts, 10 had failed at the joint, confirming our negative experience. Of 107 single-piece implants, only 11 had failed. The cause of failure in eight of 11 cases was rejection, i.e., the screws became loose from the skull. It is worth noting that six of eight cases of rejection were reported by just two laboratories. There were three cases of headpost fracture or bending. From this limited survey, it appears that most laboratories are still using acrylic to attach headposts, but a growing number of investigators are turning to single-piece, screw-mounted titanium headposts.

The surgical implantation of an acrylic-based headcap, including a chamber and headpost, can be time consuming. The acrylic must be mixed and applied to the bone in small quantities, allowing it time to harden between applications to avoid heat buildup during curing. Considerable skill is involved in sculpting a smooth cap surface with no fault lines or air inclusions. Particular care must be paid to the margins, which must be kept smooth and flush with the bone to produce a clean interface with the scalp incision. In comparison, the surgical implantation of a footed headpost can be achieved in 1 h. The short surgery causes substantially less physiological stress on the animal. The surgical incision is smaller and, except for an approximately 10-mm-diameter hole for the post, it heals closed. Implantation of a footed headpost is not difficult surgery technically.

There are considerable operational benefits to the footed titanium headpost. The most important is that the bone is able to grow and assimilate the implant. The footplate was made to be thin (1 mm) so that the bone could grow over the top of it. The enveloping of the implant in bone adds to the strength of its attachment by increasing the surface areas of screw threads and implant in contact with healthy bone tissue. Normal bone healing and growth require the intimate apposition of periosteum and bone. The periosteum contains progenitor cells that develop into osteoblasts and chondroblasts, responsible for depositing new bone tissue (Dunham 1992). The deposition of new bone tissue in response to implantation was observed after 1 mo (Fig. 5). After 14 mo, the footed implant and screws were partially embedded in the skull (Fig. 7). This process is impeded when the periosteum is stripped from the bone and the skull surface is covered by acrylic.

Various surface coatings have been used in medical implants to increase the strength of the bone–implant interface. For example, hydroxylapatite coatings have been shown to increase both the speed of formation and the strength of the adhesive bond between implant and bone (Cook et al. 1992). The main load-bearing components of our implant are the titanium screws. Most of their strength results from the mechanical apposition of the entire thread of the screw with healthy bone tissue. Adhesion between metal and bone plays a smaller role. One could apply an osteophylic coating to the implant, but this would substantially increase the cost of manufacture.

Traditional acrylic headposts are prone to infection and often require daily maintenance, including the application of antiseptic agents and antibiotics. Should a chronic infection develop, the chance of eventual failure of the headcap is increased. The smaller wound margin and smooth, metallic surface of the footed headpost mitigate this problem. The final appearance of the wound margin varied from animal to animal in our series. In some cases the scalp formed a tight interface with the post (Fig. 6) and the margin required no regular cleaning or disinfection. In others, the skin retracted 1–2 mm from the post base, leaving a small portion of the proximal footplate exposed. We cleaned this area prophylactically with a 1:1 mixture of 5% povidine-iodine and 3% hydrogen peroxide solution. Our design of the shape of the footplate minimized the separation of bone and scalp at the base of the post. This feature appeared to help the formation of a healthy soft tissue interface with the implant. We also made sure in closing the surgical wound to bring the galea over the footplate, suturing it around the base of the post to support the overlying scalp.

In addition to a headpost, awake behaving physiology experiments require the implantation of a recording chamber that allows electrodes to be inserted through a hole in the skull using a microdrive. The traditional method is to embed both the headpost and recording chamber in dental acrylic. Experiments that use search coils for measuring eye movements also require an electrical connector to be mounted to the skull. Even if acrylic is still used for the chamber and connector implantation, the size of the implant may be reduced substantially if the headpost is fixed independently without using acrylic. This is sometimes feasible, depending on the location being targeted for brain recording. We have designed and implanted a titanium recording chamber that also attaches to the skull with screws. This device promises to eliminate altogether the use of dental acrylic.


This work was supported by National Eye Institute Grant RO1 EY-10217 and Core Grant EY-02162. Monkeys were supplied by the California National Primate Research Center, supported by National Institutes of Health Division of Research Resources Base Grant RR-00169.


L. C. Sincich provided comments on the manuscript. We thank colleagues in other laboratories, especially A. F. Rossi and P. N. Sabes, who generously shared experience and expertise with the development of titanium, acrylic-free headposts. Ray Bandar lent the macaque skulls. Thanks to the veterinarians, technicians, and husbandry staff at the California National Primate Research Center and University of California San Francisco for superb care of the animals.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract