Secondary Logo

Journal Logo


Concepts of Pressure in an Ischial Containment Socket: Measurement

Neumann, Edward S. PhD, PE, CP; Wong, Jocelyn S. BS; Drollinger, Robert L. MSBE

Author Information
JPO Journal of Prosthetics and Orthotics: January 2005 - Volume 17 - Issue 1 - p 2-11
  • Free

Ischial containment transfemoral sockets can present challenging fitting problems. The fitting process requires obtaining data by several means, including questioning the patient to learn what he or she is experiencing, listening to unsolicited verbal reports of discomfort from the patient, making visual observations of tissue colors and contact pressures through the clear walls of a check socket, and using the finger or another object as a probe to estimate pressure magnitudes. Clinical tools that can improve the quality of the data obtained or enhance understanding of the phenomena that occur as a socket is loaded during gait could lead to improved socket designs and better-fitting sockets. Previous research examined the use of psychophysical scales and signal detection theory to quantify perceptions of pressure and discomfort on the residual limbs of transtibial amputees but without the availability of data on the actual pressures inside the sockets.1,2 The research reported in the current article extended the previous studies by measuring socket-limb interface pressures inside a transfemoral socket during normal gait using F-Socket (Tekscan, Inc., South Boston, MA) and comparing the magnitudes of the pressures to subjective sensations measured by psychophysical methods. This article discusses the methodology developed for obtaining and examining socket pressures, and addresses the following problems and issues: 1) To what extent can the pressures occurring inside a transfemoral ischial containment socket during gait be measured using a commercial pressure sensing system designed for prosthetic sockets? 2) How can the raw pressure data be reduced and depicted in a format that makes them amenable to a comparison with psychophysical data concerning the subjective pressures experienced by the user? 3) How well do the pressure data agree with predictions made some years ago by Radcliffe3 for quadrilateral sockets? The pressure data were used subsequently to examine perceptions of pressure. Pressures are reported in mmHg, rather than kPascals, to facilitate interpretation with respect to circulation (1 kPa = 7.5 mmHg; 1 psi = 51.71 mmHg).


Relatively few comprehensive theoretical models of socket pressure based on engineering mechanics exist to guide the prosthetist in understanding how and why pressures will vary in the different regions of the socket during gait. Perhaps the most easily comprehended and widely known of the theoretical models is the one proposed by Radcliffe3 with respect to quadrilateral sockets. This model, which emphasizes the role of the femur in maintaining anterior-posterior (A-P) and medial-lateral (M-L) stability during gait, was derived from classic engineering mechanics principles and observations made in a clinical setting, rather than actual pressure measurements. However, its predictions can serve as the basic hypothesis and starting point for interpreting and explaining pressure data obtained in the research laboratory. This approach was followed when examining the internal consistency and validity of the data obtained using F-Socket.

Generally, research studies of socket pressures have used instrumented sockets featuring force transducers. Among transfemoral socket studies, quadrilateral sockets have been examined more frequently than ischial containment sockets. Appoldt et al.4 examined both normal and shear pressures in two quadrilateral sockets during gait and found normal pressure ranges to be 258 to 1,039 mmHg latero-distal and 414 to 910 mmHg anterodistal. Upward tangential shear was found at distal locations, which led to the hypothesis that pistoning of the femur might cause sections of tissue remote from the femur to move proximally. Redhead5 proposed a hydrostatic model of tissue response in which optimum loading could be achieved using a casting procedure that involved distracting the limb by a weight equal to the weight of the prosthesis and applying a radial compression of 15.5 to 25.8 mmHg uniformly across the residual limb. Redhead proposed that this would minimize the horizontal distance from the femur to the socket interface and thereby minimize internal lost motion between the femur and socket and increase the M-L and A-P stability of the femur. Using an instrumented quadrilateral socket and taking measurements during double- and single-foot stance (not gait), highest pressures were found at the distal femur (445 mmHg), and lowest pressures were found at the lateral proximal brim (186 mmHg). The high distal pressure was attributed to the nonmyoplastic nature of the stump.

Naeff and van Pijkeren6 compared plug sockets, a pressure cast suction socket, and quadrilateral sockets made with a casting brim, including one series of quadrilateral sockets featuring different radii of curvature at the brim. The latter was tested by a nonamputee. With the quadrilateral socket series featuring varying radii (nonamputee), pressure increased with decreasing radius at the brim and had a maximum value of 7,500 mmHg for a 7.5-mm radius; this was experienced as sharp and painful by the subject. Peak pressure under the ischial tuberosity was 1,755 mmHg. For the quadrilateral sockets worn by the amputees, pressures were found to decrease gradually toward the distal regions and to peak during different phases of gait at different locations in the socket. Comparing the quadrilateral and plug sockets for a muscular residual limb and a flabby residual limb, differences were found in the magnitudes and distributions of pressure that could be attributed to both socket design and tissue stiffness. Just distal to the brim, many pressures were found to be between 150 and 750 mmHg for both types of sockets and both types of limbs.

Krouskop et al.7 compared 18 different transfemoral amputees during stance; 12 were using rigid-wall quadrilateral sockets, and 4 were using rigid-wall, normal-shape, normal-alignment (NSNA) sockets. Four quadrilateral users and one NSNA user were experiencing discomfort. With quadrilateral sockets, most of the loading occurred under the brim and varied around the circumference of the socket. The highest pressures occurred on the ischial seat and lateral wall. Pressures increased rapidly from 0 mmHg at the top of the inside to 110 to 140 mmHg at a location 5 to 8 cm from the top, then decreased over the next 5 to 10 cm to a relatively low uniform distribution. The NSNA socket produced the highest loading in the proximal one-third of the socket, but also produced similar loadings around the distal end of the femur. It was found that if the pressure in the proximal one-third was less than 95 mmHg, pistoning occurred. In the comfortable sockets, most tissues were subjected to pressures around 50 mmHg and had maximum pressures between 110 and 140 mmHg. The uncomfortable sockets generated similar pressure patterns but pressures greater than 150 mmHg. The authors concluded that pressure patterns are similar for sockets of the same design but differ for sockets of different design.

The type of instrumentation used in the previous studies is not feasible for pressure measurement in a clinical setting because holes must be drilled into the socket wall, destroying socket integrity, and creating a significantly heavier socket, which may require running numerous cables from the socket to a PC. F-Socket is a socket pressure measurement system based on the same hardware and software as the F-Scan in-shoe pressure measurement system (Tekscan, Inc.). The F-Scan sensor is shaped like the foot, consists of 960 individual cells, and is trimmed to fit the shoe of the subject. The F-Socket sensor is rectangular (21.5 × 7.5 cm), consists of 96 individual cells arranged in 16 rows and 6 columns, and has a lower spatial resolution than the in-shoe sensor (Figure 1). It is approximately 0.28 mm thick and very flexible. The sensors employ force-sensing resistors based on a piezoelectric ink sandwiched between two Mylar layers. As pressure increases, the contact area between ink particles increases, and resistance to current flows through the ink changes. The sensors must be calibrated before use, and if data are being collected for scientific research, the manufacturer recommends that the sensors also be equilibrated to account for slight variations in cell sensitivity created during manufacturing.

Figure 1.:
F-Socket and F-Scan sensors and a duplicate of the socket used in the research.

The software and hardware allow the simultaneous collection of pressure data from one foot sensor and one socket sensor. Four hundred frames of data can be collected in one trial. At a default rate of 37.9 frames per second, a frame is collected every 0.026 seconds for 10.55 seconds, which is enough time to measure approximately seven complete and sequential steps at a normal cadence. A small but long cable connects each of the two sensors to the PC used to record data, and the cables can be positioned on the subject with Velcro straps to minimize interference with gait.1 (The latest version of F-socket has eliminated the need to run cables.)

Studies of F-Scan reliability and validity have identified several characteristics that can influence data quality. Luo et al.8 examined the in-shoe sensor. Load response was found to be good when the sensor was sandwiched between two soft surfaces; slightly poorer, but still good, when sandwiched between a hard and soft surface; but off by nearly 300% when sandwiched between two hard surfaces, apparently because the hard contact surfaces do not allow sufficient flexibility of the sensor for the peaks and valleys of the ink surface to contact each other. Creep effects were slight, and dynamic load error was about 20% at 360 mmHg but decreased at higher pressures. Output was found to be sensitive to temperature and increased rapidly above 30°C. Conclusions were that accuracy worsens considerably when used on hard surfaces, and calibration must be similar to actual dynamic loading conditions. Sumiya et al.,9 using a force plate, found that the in-shoe sensor had a delayed rise time and might require as long as 1.6 seconds for creep loading to occur and stability to appear in measurements. They also found large errors when loads were applied rapidly to sensors on hard surfaces and concluded that the structure of the sensor, which had sensing areas 0.02 to 0.04 mm thicker than the nonsensing areas, led to a concentration of load on the sensing areas, and a higher output. The authors also stated that the sensor detects shear in addition to normal forces, so measurements reflect both. They concluded that while the F-scan might not measure normal pressures accurately enough for a high level of certainty with respect to absolute values, it does allow relative comparisons of pressure distributions and is valid for the evaluation of footwear fit for individuals who have normal sensations.

Polliack and colleagues10,11 compared a variety of socket pressure sensing systems and used both a flatbed and the mold of a transtibial residual limb enclosed in a pressure chamber. Readings were taken in 517-mmHg increments from 517 to 2,587 mmHg. For the mold, the accuracy error was found to be roughly 11%, hysteresis error was 24%, and drift error was 33%. The accuracy error was least near the calibration pressure of 2,068 mmHg. Readings of approximately 100 mmHg could be produced simply by flexing the sensor. The authors favored the F-Socket over the other systems but found that accuracy worsened with pressures greater than those used to calibrate the sensor. They concluded that the large hysteresis and drift errors would affect measurements during gait, when a socket is cyclically loaded and unloaded, and sessions may last longer than 20 minutes. They recommended that the F-socket be used only in a static or standing environment, where a need exists to observe pressure gradients or areas of high pressure.

F-Scan has been used in numerous studies of plantar pressure, but reported studies of socket pressures using F-Socket are relatively few. Convery and Buis12 used F-Socket to measure dynamic pressure distributions during prosthetic stance for a conventional patellar-tendon bearing socket. Four separate sensors were attached to the anterior, posterior, medial, and lateral walls using adhesive, and calibration was accomplished by inserting a gel-filled sheath and creating a known pressure. Data were collected for two sensors at a time and then matched. A force plate was used to measure ground reaction force and gait parameters. Subsequently, Convery and Buis13 used F-socket and the same sensor attachment and calibration technique to compare pressure distributions for both standard PTB and pressure cast transtibial sockets.


A subject was recruited who had, because of trauma, sustained a transfemoral amputation just proximal to the femoral condyles on the left side and had been fitted with an ischial containment suction socket featuring a flexible inner socket and frame with windows. The subject’s prosthesis featured a hydraulic single-axis knee, an energy storing foot, and a torque absorber. The subject was generally pleased with the fit of the socket, which had been worn for 6 years and was reported as being snug but comfortable, although pressures capable of traumatizing the subject’s tissues were experienced occasionally at the cut end of the femur on the lateral side. The subject’s femur was 22 cm long measured from the perineum and featured a moderate amount of unanchored loose tissue with considerable longitudinal and rotational mobility with respect to the femur. Surgery appeared to have been nonmyodesic, and the myoplasty that was performed left little tissue between the skin and the cut end of the femur on the lateral side. Evaluation of the subject’s sensitivity to touch by means of a biothesiometer revealed no impairment of sensation. The subject reported preferring to don the socket by pulling into it, rather than using lotions, because the quality of the fit seemed better. The subject was very active and used a prosthesis while swimming, cycling, hiking, and scuba diving. There were no reported problems with loss of suction, and the alignment of the prosthesis and gait of the subject appeared very good. The subject read and signed a letter of informed consent that had been approved, along with the human subject protocol, by the Biomedical Sciences Committee of the Institutional Review Board of the University of Nevada, Las Vegas.

The sensor was first equilibrated according to the manufacturer’s recommendations. Next, calibration was performed by sandwiching the sensor between two layers of 14-mm thick Plastazote (American Plastics, Forth Worth, TX) cut to the exact dimensions of the sensor to prevent possible bridging. The Plastazote layers in turn were sandwiched between two layers of 9-mm thick clear co-polyester to provide stiffness. The subject was asked to stand on the sandwich of Plastazote, clear co-polyester, and sensor during calibration, and the subject’s weight entered into the software, which performed the calibration. The in-shoe sensor was trimmed to fit inside the shoe on the prosthetic side, equilibrated, and calibrated following the manufacturer’s recommendations, and inserted between the prosthetic foot and the shoe.

The socket sensor was positioned inside the socket beginning at the anterior over Scarpa’s triangle, and the socket was donned using a commercial parachute silk device. After the subject determined that the socket had been donned correctly and felt “right,” data collection was initiated. The subject was asked to walk in a straight line at a comfortable, normal pace until 400 frames had been captured. Once a successful trial had been captured, the subject was asked to identify, for each of the muscle compartments and high pressure regions that were covered by the sensor, the phases of stance when both maximum and minimum pressures were experienced, and the subjective magnitude of the pressure. A psychophysical scale was used to report the amount of subjective pressure experienced. The location of the midpoint and edges of the sensor were marked on the socket, which was then doffed, and the sensor was repositioned in a counterclockwise direction looking downward into the socket (i.e., from Scarpa’s triangle toward the lateral wall) to capture pressures in the region immediately adjacent to the previous region. The socket was then donned again. These steps were repeated until pressure data for the entire inside of the socket had been obtained to the depth of the sensor length, which resulted in seven different locations. Attempts were made to position the sensor about the same distance from the distal end of the socket but high enough to capture important high-pressure regions around the ischium and ramus; however, this resulted in not completely capturing the region around the femoral relief, so an eighth set of pressure measurements was taken by placing the distal end of the sensor at the very bottom of the socket over the cut end of the femur on the lateral side. Because the segment lengths of the individual’s thighs were relatively short, the sensors reached all but the most distal 3 or 4 cm of the socket.

Figure 2 (top) indicates the typical appearance of the pressure measurements, using the contour display feature, for one frame of data for each of the eight different sensor locations at approximately the same point in the gait cycle during loading response. To analyze pressure, areas of the sensors were windowed as shown in the bottom of Figure 2, and the average pressures over all the cells in each window were displayed. The windows were defined to represent the proximal and distal compartments for major muscle groups and smaller areas where socket fit can be troublesome because of bony prominences, tendons, or muscle contractions. Four phases in the gait cycle corresponding to loading response, midstance, terminal stance, and midswing were identified from an examination of the plantar pressure patterns of the in-shoe sensor. With the use of Tekscan software (Version 4.21), the frames of the movie containing peak average window pressures were identified for each phase of gait over five consecutive steps beginning with the second step. The average pressures associated with each window were recorded, and the means and standard deviations of these values for the five steps were calculated.

Figure 2.:
(Top) F-Socket pressure contour maps. Blue spans 0 to 180 mmHg in 60 mmHg increments, green spans 180 to 270 mmHg, yellow and orange span 270 to 360 mmHg, and red represents pressures of 360 mmHg and higher. (Bottom) Corresponding windows used for obtaining the means of the average pressures for muscle compartments and regions of high pressure. Windows for the muscle compartments were defined as follows: knee extensors = anterior and antero-lateral sensors; hip abductors = lateral sensor; hip extensors = postero-lateral and posterior sensors; and hip adductors = adductor magnus and medial sensors. Windows for pressure sensitive areas were defined for Scarpa’s triangle (anterior sensor), the femoral relief (lateral sensor), the gluteal fold (postero-lateral and posterior sensors), the ischium (posterior sensor), the ramus (medial sensor), and the adductor longus tendon (medial sensor).

Polar coordinates were then used to plot the mean of the average pressures over these five steps as a function of the window location along the inner circumference of the socket. For the proximal and distal muscle compartments, the mean average pressures were plotted at the sensor midpoints, and the resulting points were connected with splines to produce smooth curves using SYSTAT 8.0 (SPSS, Inc., Chicago, IL). The locations of the resulting sensors on the plots corresponded well, but not exactly, with the actual locations of the sensors when looking down into the socket.

For the smaller regions of potentially high pressure, the angular location for each window midpoint was adjusted on the plots to reflect the actual location of the window with respect to the midpoint of the sensor. The numbers 1, 2, 3, and 4 were plotted to represent the phase of gait corresponding to loading response, midstance, terminal stance, and midswing, respectively; and the letters A through G were located along the outer edge of the plot to identify the socket feature.


Figures 3A and B indicate the pressure maps for the proximal and distal muscle compartments by phase of gait, and Table 1 presents the values. Mean measured pressures between 100 and 300 mmHg were obtained for virtually all muscle compartments and all phases of gait. The highest pressures for all the muscle compartments except the distal anterior were associated with loading response. In all compartments except the distal and proximal anterior, pressures then decreased toward midstance and in some cases exhibited small increases at terminal stance. The distal anterior compartment exhibited a maximum pressure at midstance, and the proximal anterior compartment exhibited approximately equal pressures at loading response and midstance. The decreases in mean pressure during gait generally were greatest in the distal compartments. Figures 3C–F compare mean proximal and distal muscle compartment pressures for each of the four phases of gait. It can be seen that the pressure patterns vary substantially between the distal and proximal portions of the socket, indicating the existence of gradients. During all phases of stance, the distal compartments exhibited higher pressures than did the proximal compartments in the antero-lateral, lateral, and postero-lateral regions, whereas the proximal compartments exhibited higher pressures than did the distal compartments in the anterior, medial, and adductor magnus compartments. In the posterior compartment, the pattern changed from being highest distally during loading response to being highest proximally during terminal stance.

Figure 3.:
Pressure maps showing mean average pressures for muscle compartments. The x-axis ranges from 0 to 300 mmHg. Pressures were plotted at the midpoints of each window (each of the seven sensors spans 360/7 = 51.4°). Looking down into the socket, a line approximately perpendicular to the line of progression during gait was selected as 0°, and the clockwise edge of the medial sensor containing the adductor longus tendon is located at 0°. The midpoint of the first sensor is 25.7° counterclockwise, and the midpoint of each successive sensor is located 51.4° counterclockwise from the midpoint of the previous sensor. LR = loading response; MST = midstance; TS = terminal stance; MSW = midswing.
Table 1:
Pressures in muscle compartments and regions of high pressure, averaged over five steps during gait

Figures 4A and B are plots of the recorded mean average and mean peak pressures, respectively, for the areas of potentially high pressure. Mean average pressures as high as 600 mmHg were recorded in the adductor longus, ramus, and ischial areas of the socket near the brim, nearly double the pressures in the muscle compartments. Mean peak pressures as high as nearly 1,200 mmHg were recorded for the ramus region during midstance and approximately 1,000 mmHg at the femoral relief on the lateral side during the same phase of gait. While folding of the sensor at the brim of the socket could cause elevated pressures in a few cells, it is known from fitting experience that these locations are a common source of discomfort problems. The patterns of pressure change for the ramus and distal femur were highly similar, reaching a maximum at midstance, a minimum at terminal stance, and an intermediate value at loading response, suggesting a possible link between the two. Pressure at Scarpa’s triangle also tended to peak toward midstance and had a pattern of change similar to the distal femur and the ramus. Pressure patterns at the ischium and the gluteal fold appeared related, having a minimum during loading response followed by an increase during midstance and a peak during terminal stance. Pressures in the region of the adductor longus tendon also peaked at terminal stance.

Figure 4.:
Pressure maps for regions of high pressure. The x-axis ranges from 0 to 1,200 mmHg. LR = loading response; MST = midstance; TS = terminal stance; MSW = midswing. Letters A through G represent mean average pressures (a) or mean peak pressures (b) for the following windows: SCARPA = Scarpa’s triangle; FEMUR = femoral relief; GLUTE1 = postero-lateral gluteal fold; GLUTE2 = posterior gluteal fold; ISCHIUM = ischium; RAMUS = ramus; ADLONGT = adductor longus tendon.

The pressure variations during the five steps used to calculate means are expressed as coefficients of variation (standard deviation divided by the mean pressure) in Table 1. In all the muscle compartments except the proximal adductor magnus and the distal lateral, the highest relative step-to-step pressure variation occurred either during loading response or terminal stance, and the least during midstance. In the proximal adductor magnus and distal lateral compartments, the maximum coefficients of variation occurred during midstance. In the high-pressure regions, the highest coefficients of variation occurred during loading response in all but Scarpa’s triangle and the ramus, where the maximum occurred during terminal stance. The highest coefficients of variation for all the windows occurred in the femoral relief during loading response and midstance, which as noted previously, was a location of potential discomfort for the subject, and in the proximal adductor magnus compartment during midstance.


The data reduction techniques involved some subjectivity, and it was not possible to take all measurements simultaneously, which would have ensured that data were extracted from all movies at exactly the same point during gait. However, the measured pressures appeared to have internal consistency in terms of magnitudes and gradients. There were no unusual spikes or regions with unexpected pressure patterns, and regions known to create high pressures inside the socket also exhibited high pressures on the sensors. Even if recorded pressures were either higher or lower than actual pressures because of drift, delayed response, or calibration, the relative pressures could be viewed as valid, particularly for the muscle compartments where no folding of the sensors occurred. The pressure values obtained for the muscle compartments were within the ranges reported in all the previously cited studies, and in many cases lower, with one exception, which were the pressures measured by Krouskop et al.7 The values the Krouskop group cited as the upper boundary for a comfortable socket were roughly 100 to 150 mmHg less. Judging from the fact that pressures during midstance were often greater than 100 mmHg in most of the muscle compartments, it is possible that drift was occurring because of prestressing of the sensors subsequent to donning but before data collection. If this were the case, the actual pressures could be closer in value to those obtained by Krouskop et al.

Variability of movement is an inherent characteristic of motor control, and it may be related to stability, skill, or adaptability to varying task demands.14 The phases of stance associated with the highest coefficients of variation may have been related to knee stability during loading response and terminal stance, and pelvic stability during midstance, as well as the skill with which the subject was able to control the prosthesis. With respect to adaptability, the high coefficient of variation associated with the femoral relief, ramus, and adductor magnus compartments also may have been related to the greater pressure on the tissues near the cut end of the femur and a desire on the part of the subject to avoid discomfort.

Radcliffe’s hypothesized pressure relationship for quadrilateral sockets states that forces during gait act in the A-P direction to stabilize the knee and in the M-L direction to stabilize the pelvis.3 To stabilize the pelvis in the M-L direction, Radcliffe’s hypothesis predicts that a force couple produced by the femur will create the highest pressures in the proximal medial and distal lateral compartments during midstance. The data obtained were in agreement with this prediction with respect to the ramus and femoral relief. However, the pressures in the adjacent muscle compartments, proximal medial and distal lateral, were highest during loading response. Concerning knee stability and forces in the A-P direction, Radcliffe’s hypothesis predicts that the highest pressures during loading response will occur in the proximal anterior and distal posterior compartments, and the highest pressures during terminal stance will occur in the distal anterior and proximal posterior compartments. The data obtained were in agreement with this hypothesis during loading response for the proximal anterior and distal posterior compartments. But the distal anterior compartment and proximal posterior compartments had similar pressures for all three phases of stance and were not in agreement with the hypothesis. But the data for the gluteal fold and ischium were in agreement with Radcliffe, implying that the high forces were in the predicted locations but concentrated over smaller regions of the socket. A third departure from Radcliffe’s theory was found at Scarpa’s triangle, where Radcliffe’s hypothesis predicts that highest pressure will occur during loading response. The data indicated that highest pressure in Scarpa’s triangle tended to occur toward midstance, and the pattern was similar to the one associated with M-L pelvic stability. A possible explanation is that because the socket was flexible, it tended to expand in the M-L direction in response to the force couple created by abduction of the femur during midstance, and this caused a contraction in the A-P direction, increasing pressure at Scarpa’s triangle.

The data suggest that although Radcliffe’s hypothesis appeared to be in general agreement, the highest pressures tended to concentrate on small regions of the socket where the subject’s limb may have offered the greatest stiffness, rather than the muscle compartments. It is possible that the prosthesis with socket and the mobile tissues of the muscle compartment created what could be considered a “wobbly” unit of mass that had its movement constrained by the smaller regions of higher pressure, most of which were created by the skeleton and tendons. This wobbly unit of mass might serve somewhat like a shock absorber.15 It can be hypothesized that with a different socket rectification, distributions could be altered to reduce peak and average pressures in the high pressure areas and transfer more of the loads to the larger volumes of tissue in the muscle compartments.

Radcliffe’s hypothesis concerns quadrilateral sockets for which the ischial seat helps constrain downward movement of the pelvis and the femur. The effects of downward movement of the femur on pressure distributions in the distal compartments of an ischial containment socket might be greater and would be dependent on the biomechanical properties of the tissues around the femur, as well as the extent to which the tissue layers are attached to each other and to the femur and resist movement with respect to each other and the femur. It is highly probable that the femur moves distally into an ischial containment socket during loading response, especially if the socket is flexible, and creates stresses that the surrounding tissues must absorb. The femur also abducts, adducts, extends, and flexes with respect to the socket wall.16,17 These movements produce stress on the adjacent tissues, and the amount of tissue deformation and displacement are governed by the extent to which the tissues resist normal and shear strain.18–22

Concepts of normal and shear strain can be applied at the macro level of limb morphology to propose a heuristic model that can help to visualize the effects of femur pistoning on the pressure patterns in the distal socket of the subject. Tissues of the limb are usually assumed to be incompressible, the most probable scenario for very rapid, short-term loadings of the type experienced during gait when fluids in the circulatory system and interstitial fluids probably do not have sufficient time to flow to regions of lower pressure. The tissues of the subject were mobile in the distal portion of the limb, had a high proportion of adipose cells, likely had little anchoring to muscle or the femur, and offered little resistance to shear-type movements both longitudinally and rotationally with respect to the femur. In this respect the distal tissues behaved somewhat like a collection of small, incompressible fluid-filled sacs that were connected to one another by common walls and enclosed in an outer cover of skin, but because of their minimal resistance to shear forces, permitted considerable change in the geometry of their overall envelope. Because of this and the fact that the socket was snug, the tissues behaved in a manner that has been termed “quasi-hydrostatic” by some prosthetists, as discussed in greater depth in Klasson.23 At more proximal locations on the femur, there was more anchoring of the underlying muscles to the femur, which created a cone of more shear-resistant tissues attached to the femur. This cone of more shear-resistant tissues was surrounded by the envelope of less shear-resistant adipose tissue. As the femur displaced distally, the less easily deformable cone pressed into the more easily deformable surrounding envelope of adipose tissue, which was constrained by the socket wall and responded somewhat like a fluid-filled bag, creating relatively smooth pressure gradients in the distal portion of the socket. At more proximal locations in the socket, the response of the more shear-resistant, less mobile tissues created a more sculpted pressure pattern, the shape of which was influenced by the cross-sectional design of the socket.

The pressure patterns suggest that among patients having a limb morphology similar to that of the subject’s, with mobile adipose tissues, downward displacement of the femur after loading response could increase the pressure between the distal limb and the socket, increasing the stiffness of the coupling and the proportion of gait-related energy that is absorbed by the distal tissues. With careful fitting techniques that ensure a snug socket, it might be possible to take advantage of this and thereby reduce pressures in more sensitive proximal regions. A loosely fitting socket would not be able to develop quasi-hydrostatic loading in the distal tissues, and the forces would be transmitted to other regions and tissues of the limb based on their relative stiffness. However, the coefficient of friction between the skin and socket could influence the possible benefits of quasi-hydrostatic loading because normal pressure would decrease as frictional resistance between the skin and the socket increased, and with increasing frictional resistance of the skin, internal stresses within the tissues that lie between the skin and the femur might also increase.24


Pressure mapping of the type undertaken in this study facilitates visualization and interpretation of the dynamic interplay between the temporal variation in ground reaction forces and socket pressures at critical locations inside the socket. The measured pressures were in general agreement with Radcliffe’s model for A-P stability of the knee and M-L stability of the pelvis, but the data suggest that in a snugly fitting socket, downward movement of the femur may create additional quasi-hydrostatic pressure, which is superimposed on the pressure related to knee and pelvic stability. A heuristic model was proposed to help visualize the morphology of the limb of the subject: a cone of more shear-resistant proximal tissues attached to the femur, which transitioned into a cylinder of less-shear resistant adipose distal tissues, and together with the femur, contributes to the generation of quasi-hydrostatic pressure in a snug socket. However, these findings are tentative and the hypotheses intended to help other researchers. It should be noted that pressures related to the rotational torque experienced during gait, which would create a force moment in the transverse plane of the prosthesis and socket, might also be present in the data. A force plate or additional instrumentation of the prosthesis would be required to determine this.

Pressure measurement of an entire socket using the methods of this study is time consuming, but the results produce a valuable educational tool. Clinical application of these methods might be more limited. Data collection for one subject can easily take 2 hours, which is more than some patients would be willing to spend. Use of multiple precalibrated sensor strips would facilitate simultaneous collection of data at all points in the socket but would be extremely expensive and require multiple hardware set-ups because F-Socket can record pressure from only two sensors at one time. Thus, it is unlikely that a practitioner in a clinical setting routinely would want to map an entire socket unless a patient is experiencing a discomfort problem that is resistant to resolution, and it is impossible to observe what is occurring inside the socket (e.g., a laminated socket). However, pressure mapping of regions of a socket might be helpful in situations in which a need exists to establish whether a discomfort problem can or cannot be resolved by socket modification. Pressure measurement together with subjective measurement using psychophysical scales could be useful for diagnosing troublesome socket discomfort or tissue trauma problems, particularly for patients whose nerves have become sensitized by surgical procedures or underlying pathologies.

A next step would be to develop visualization tools that would allow the prosthetist to overlay pressure data from currently available two-dimensional sensing systems onto three-dimensional images of the limb. If F-Socket is used, additional work needs to be undertaken to identify the best methods for calibrating the sensors. Given the potential influences that temperature, drift, hysteresis, and the hardness of the calibrating surface can have on the output from the sensors, research should be undertaken to determine the magnitude of the potential error and develop calibration methods that minimize error or correction factors.

Socket rectification procedures typically are conceptualized in terms of sculpting templates, and many different normative designs have appeared. The biomechanical rationales underlying these templates are rarely examined and validated using concepts from engineering mechanics. With improved methods for characterizing limb morphology, and measuring, depicting, and analyzing socket interface pressures, knowledge of how the musculoskeletal system and tissues of the residual limb respond to these designs during gait can be advanced. This advancement in knowledge would lead to improved designs, better-tailored to the unique needs of each individual.


1. Neumann ES. Measurement of socket discomfort. Part I: pressure sensation. J Prosthet Orthot 2001;13:99–110.
2. Neumann ES. Measurement of socket discomfort. Part II: signal detection. J Prosthet Orthot 2001;13:111–122.
3. Radcliffe CW. Prosthetics. In: Rose J, Gamble JG, eds. Human Walking. Baltimore: Williams & Wilkins; 1994:168–173.
4. Appoldt FA, Bennett L, Contini R. Tangential pressure measurements in above-knee suction sockets. Bull Prosthet Res 1970;10–13:70–86.
5. Redhead RG. Total surface bearing self suspending above-knee sockets. Prosthet Orthot Int 1979;3:126–136.
6. Naeff M, van Pijkeren T. Dynamic pressure measurements at the interface between the limb and socket: the relationship between pressure distribution, comfort, and brim shape. Bull Pros Res 1980;10–17:35–50.
7. Krouskop TA, Brown J, Goode B, Winningham D. Interface pressures in above-knee sockets. Arch Phys Med Rehabil 1987;68:713–714.
8. Luo ZP, Berglund LJ, An KN. Validation of F-scan pressure sensor system: a technical note. J Rehabil Res Dev 1998;35:186–191.
9. Sumiya T, Suzuki Y, Kasahara T, Ogata H. Sensing stability and dynamic response of the F-scan in-shoe sensing system: a technical note. J Rehabil Res Dev 1998;35:192–200.
10. Polliack A, Landsberger S, McNeil D, et al. Socket measurement systems perform under pressure. Biomechanics 1999;6(June):71–80.
11. Polliack AA, Sieh RC, Craig DD, et al. Scientific validation of two commercial pressure sensor systems for prosthetic socket fit. Prosthet Orthot Int 2000;24:63–73.
12. Convery P, Buis AWP. Conventional patellar-tendon-bearing (PTB) socket/stump interface dynamic pressure distributions recorded during the prosthetic stance phase of gait of a transtibial amputee. Prosthet Orthot Int 1998;22:193–198.
13. Convery P, Buis AWP. Socket/stump interface pressure dynamic pressure distributions recorded during the prosthetic stance phase of gait of a trans-tibial amputee wearing a Hydro-Cast socket. Prosthet Orthot Int 1999;23:107–112.
14. Newell KM, Corcos DM, eds. Variability and Motor Control. Champaign, IL: Human Kinetics, 1993.
15. Nigg BM. Mathematically Determinate Systems. In: Nigg BM, Herzog W, eds. Biomechanics of the Musculoskeletal System. Chichester, United Kingdom: John Wiley & Sons, 1999:525–532.
16. Convery P, Murray KD. Ultrasound study of the motion of the residual femur within a transfemoral socket during gait. Prosthet Orthot Int 2000;24:226–232.
17. Convery P, Murray KD. Ultrasound study of the motion of the residual femur within a transfemoral socket during daily living activities other than gait. Prosthet Orthot Int 2001;25:220–227.
18. Vannah WM, Childress DS. Modeling the mechanics of narrowly constrained soft tissues: the effects of specification of Poisson's ratio. J Rehabil Res Dev 1993;30:205–209.
19. Zachariah SG, Sanders JE. Interface mechanics in lower-limb external prosthetics: a review of finite element models. IEEE Trans Rehab Eng 1996;4:288–302.
20. Silver-Thorn MB, Childress DS. Generic geometric finite element analysis of the transtibial residual limb and prosthetic socket. J Rehabil Res Dev 1997;34:171–186.
21. Mak AF, Liu GH, Lee SY. Biomechanical assessment of below-knee residual limb tissue. J Rehabil Res Dev 1994;31:188–198.
22. Zheng YP, Mak AF, Leung AK. State-of-the-art methods for geometric and biomechanical assessments of residual limbs: a review. J Rehabil Res Dev 2001;38:487–504.
23. Klasson B. Appreciation of Prosthetic Socket Fitting From Basic Engineering Principles. Technical report WE 172 KLA. Glasgow: National Centre for Training and Education in Prosthetics and Orthotics, University of Strathclyde, 1995.
24. Zhang M, Mak FT. A finite element analysis of the load transfer between an above-knee residual limb and its prosthetic socket: roles of interface friction and distal-end boundary conditions. IEEE Trans Rehab Eng 1996;4:337–346.

dynamic socket pressure; ischial containment socket; pressure measurement

© 2005 American Academy of Orthotists & Prosthetists