Journal of Electromyography and Kinesiology 10 (2000) 361–374
`
`www.elsevier.com/locate/jelekin
`
`Development of recommendations for SEMG sensors and sensor
`placement procedures
`
`Hermie J. Hermens a,*, Bart Freriks a, Catherine Disselhorst-Klug b, Gu¨nter Rau b
`a Roessingh Research and Development, Roessinghbleekweg 33, 7522 AH Enschede, The Netherlands
`b Helmholtz Institute, Aachen, Germany
`
`Abstract
`
`The knowledge of surface electromyography (SEMG) and the number of applications have increased considerably during the
`past ten years. However, most methodological developments have taken place locally, resulting in different methodologies among
`the different groups of users.
`A specific objective of the European concerted action SENIAM (surface EMG for a non-invasive assessment of muscles) was,
`besides creating more collaboration among the various European groups, to develop recommendations on sensors, sensor placement,
`signal processing and modeling. This paper will present the process and the results of the development of the recommendations
`for the SEMG sensors and sensor placement procedures.
`Execution of the SENIAM sensor tasks, in the period 1996–1999, has been handled in a number of partly parallel and partly
`sequential activities. A literature scan was carried out on the use of sensors and sensor placement procedures in European labora-
`tories. In total, 144 peer-reviewed papers were scanned on the applied SEMG sensor properties and sensor placement procedures.
`This showed a large variability of methodology as well as a rather insufficient description. A special workshop provided an overview
`on the scientific and clinical knowledge of the effects of sensor properties and sensor placement procedures on the SEMG character-
`istics.
`Based on the inventory, the results of the topical workshop and generally accepted state-of-the-art knowledge, a first proposal
`for sensors and sensor placement procedures was defined. Besides containing a general procedure and recommendations for sensor
`placement, this was worked out in detail for 27 different muscles. This proposal was evaluated in several European laboratories
`with respect to technical and practical aspects and also sent to all members of the SENIAM club (.100 members) together with
`a questionnaire to obtain their comments. Based on this evaluation the final recommendations of SENIAM were made and published
`(SENIAM 8: European recommendations for surface electromyography, 1999), both as a booklet and as a CD-ROM. In this way
`a common body of knowledge has been created on SEMG sensors and sensor placement properties as well as practical guidelines
`for the proper use of SEMG. (cid:211)
`2000 Elsevier Science Ltd. All rights reserved.
`
`Keywords: Surface EMG; Sensor; Electrodes
`
`1. Introduction
`
`The knowledge of surface electromyography (SEMG)
`has increased considerably during the past ten years.
`This concerns a better understanding of the physiological
`processes that contribute to the generation of this signal,
`more adequate signal processing techniques and a grow-
`ing knowledge on how it can be applied in various clini-
`cal applications. In particular, the rapid growth of the
`number of applications underlines the high potential of
`SEMG as a non-invasive tool for the assessment of the
`
`neuromuscular system. On the other hand, however,
`most methodological developments have taken place
`locally, resulting in different methodologies among the
`different groups of users. This hinders the further growth
`of SEMG into a mature well-accepted tool by the users
`as well as industrial efforts on a large scale. A stan-
`dardization effort is required to make the results more
`comparable and to create a large common body of
`knowledge on the use of SEMG in the various fields
`of application.
`the European concerted action
`With this in mind,
`SENIAM (surface EMG for a non-invasive assessment
`of muscles) was started in 1996. Besides having the gen-
`eral goal of creating more collaboration among the vari-
`
`* Corresponding author.
`
`1050-6411/00/$ - see front matter (cid:211)
`PII: S 1 0 5 0 - 64 11 ( 0 0 ) 0 0 02 7- 4
`
`2000 Elsevier Science Ltd. All rights reserved.
`
`Petitioner - Avation Medical, Inc.
`Ex. 1028, p. 361
`
`

`

`362
`
`H.J. Hermens et al. / Journal of Electromyography and Kinesiology 10 (2000) 361–374
`
`ous European groups [14,15,17,19], the specific goal was
`formulated to develop recommendations on key items to
`enable a more useful exchange of data obtained with
`SEMG, including sensors, sensor placement, signal pro-
`cessing [20] and modeling [18]. Two of these key items
`involved sensors and the placement of sensors on the
`muscle. In this context
`the sensor is defined as the
`arrangement of electrodes put on the skin surface to pick
`up the EMG signal from the underlying muscle. As it is
`clear that these two items are very much interrelated it
`was decided to combine them into one set of sensor
`tasks.
`This paper will present the process and the results of
`the development of the recommendations for the SEMG
`sensors and sensor placement procedures.
`
`2. Methods
`
`Execution of the SENIAM sensor tasks, in the period
`1996–1999, has been handled in a number of partly par-
`allel and partly sequential activities. The interactions
`between these activities are shown in Fig. 1.
`First, an inventory was carried out on the use of sen-
`sors and sensor placement procedures in European lab-
`oratories. The inventory consisted of a questionnaire cir-
`culated among the SENIAM partners and a literature
`scan of 144 SEMG publications by European authors.
`In parallel, an overview was obtained on the scientific
`and clinical knowledge of the effects of sensor properties
`and sensor placement procedures on SEMG signal
`characteristics. This was done by organizing a topical
`workshop, with experts in this field, to discuss these
`various effects and to produce a consensus on relevant
`guidelines. In addition, some specific experimental stud-
`ies have been carried out in European laboratories com-
`bining the knowledge and facilities of the partners [19].
`Based on the inventory,
`the results of the topical
`workshop and generally accepted state-of-the-art knowl-
`edge, a first proposal for sensors and sensor placement
`procedures was defined [16]. Besides containing a gen-
`
`Fig. 1. Sequence and interrelations between the SENIAM sensor
`tasks.
`
`eral procedure and recommendations for sensor place-
`ment, this was worked out in detail for 27 different
`muscles. This proposal was evaluated in several Euro-
`pean laboratories with respect to technical and practical
`aspects. The proposal was also sent to all members of
`the SENIAM club (.100 members) together with a
`questionnaire to obtain their comments. Based on this
`evaluation the final recommendations of SENIAM were
`made and published, both as a booklet [21] and as a CD-
`ROM [10]. In this way a European common body of
`knowledge has been created on SEMG sensors and
`sensor placement procedures as well as practical guide-
`lines for applications.
`
`2.1. The inventory
`
`The inventory on sensors and sensor placement pro-
`cedures consisted of two parts:
`
`1. A questionnaire among the 16 SENIAM partners;
`2. A literature scan of a large number of European publi-
`cations on SEMG.
`
`The questionnaire should be regarded as a pilot study
`for the literature scan. It was designed to obtain a first
`impression about
`the sensors, sensor placement pro-
`cedures and equipment used in the European labora-
`tories. It was sent to the 16 SENIAM partners and they
`all returned the form. A main conclusion [9] was that a
`large variety of sensors and equipment is being used in
`these laboratories. The high variability in this limited
`amount of data justified a larger-scale effort.
`The literature scan was based on seven journals in
`which publications about SEMG can be found regularly.
`Table 1 shows an overview of the selected journals. The
`selection covers most of the application areas of SEMG
`as well as the more basic research-related activities. In
`the available volumes of the last 5–7 years (1991–1997)
`all publications from European first authors have been
`scanned with respect to the following subjects:
`
`1. General: author, title of the publication, journal, vol-
`ume.
`2. Sensor properties: manufacturer type, number of con-
`tact points, shape, size, material, inter-electrode dis-
`tance.
`3. Sensor placement procedure: skin preparation tech-
`nique, paste, muscles, location on muscles, location
`of reference electrode.
`4. Equipment; Signal processing; Comments.
`
`All data were entered in a database and then checked
`for completeness. In case of incompleteness, the captured
`information was put in a form and then sent to the first
`author with a request to complete the information. The
`additional information was then entered into the database.
`
`Petitioner - Avation Medical, Inc.
`Ex. 1028, p. 362
`
`

`

`H.J. Hermens et al. / Journal of Electromyography and Kinesiology 10 (2000) 361–374
`
`363
`
`Table 1
`Numbers and years of SEMG publications scanned for the inventory
`
`Journal
`
`Scanned volumes
`
`The Journal of Electromyography and Kinesiology
`Electromyography in Clinical Neurophysiology
`Electroencephalography in Clinical Neurophysiology
`The Journal of Biomechanics
`Ergonomics
`Muscle and Nerve
`The European Journal of Applied Physiology
`
`1991, 1992, 1993, 1994, 1995, 1996
`1993, 1995, 1996
`1992, 1993, 1994, 1995, 1996
`1992, 1993, 1994, 1995, 1996, 1997
`1994
`1992, 1993, 1994, 1996
`1995, 1996
`
`Number of
`publications
`
`34
`20
`38
`13
`6
`9
`24
`Total: 144
`
`3. Results
`
`3.1. The inventory
`
`3.1.1. Number of papers scanned and verified by first
`authors
`In total, 144 peer-reviewed papers were scanned. The
`number of publications on SEMG that was found in each
`journal is shown in Table 1. This table also shows which
`volumes of the journals have been scanned. Because not
`all volumes were available in the libraries visited, it was
`not possible to scan at least five complete volumes of
`each journal.
`In 101 of the 144 papers the address of the author was
`known so a request for completion could be sent. Of
`these, 33 (32%) were returned and the information
`obtained was added to the database.
`
`3.1.2. Sensor configuration
`Initially, in 40% of the publications the applied sensor
`configuration was not mentioned properly. After feed-
`back from the authors the sensor configuration in 126
`(88%) of the publications was known. This resulted in
`the following overview:
`
`O monopolar reported in 5 publications
`O bipolar reported in 115 publications
`O array/line electrodes reported in 6 publications
`
`It is very clear that a bipolar sensor configuration was
`used most frequently. Most references to monopolar
`configurations were found in ElectroEncephaloGraphy
`and Clinical Neurophysiology while most array/line con-
`figurations were found in the Journal of Electromyogra-
`phy and Kinesiology.
`
`3.1.3. Trademark
`In 81 of the 144 publications (56%) the trademark of
`the electrodes used was not mentioned. After feedback
`from the authors this amount reduced to 63 (44%) which
`means that the trademark of the electrodes in 81 (56%)
`publications was known. These were:
`
`O Medicotest (child ECG-electrodes) reported in 18
`publications
`O Beckmann (miniature size) reported in 14 publi-
`cations
`O Own/custom-made reported in 8 publications
`O DISA 13Lxx reported in 5 publications
`O Meditrace reported in 4 publications
`O Dantec reported in 4 publications
`O Other reported in 25 publications
`
`The inventory shows a large variety in electrodes used:
`in total, 24 trademarks have been counted (the 6 trade-
`marks mentioned above together with 18 other trade-
`marks which have been counted less than four times and
`make up the ‘other’ category). This variety is even larger
`when taking into account that within trademark categor-
`ies a large variety in shape and size can be discerned.
`The inventory shows that there is a preference for
`Medicotest and Beckmann electrodes, which is not sur-
`prising as they are easily available and are small enough
`to be used for SEMG recordings.
`
`3.1.4. Material
`In 82 (57%) of the publications the material of which
`the electrode was made was not mentioned. After feed-
`back from the authors this amount has been reduced to
`62 (43%). In the remaining 82 (57%) publications the
`electrodes were made of the following material:
`
`O Ag/AgCl reported in 57 publications
`O Ag reported in 11 publications
`O AgCl reported in 6 publications
`O Au reported in 3 publications
`O Other materials reported in 5 publications
`
`In total, seven types of electrode materials have been
`discerned (the four types mentioned plus three types in
`the ‘other materials’ category: tin, metal, stainless steel,
`each mentioned less then three times). For bipolar or
`monopolar electrodes, it is obvious that Ag/AgCl was
`the preferred electrode material. Although reported sep-
`arately, it makes no sense to distinguish between AgCl
`
`Petitioner - Avation Medical, Inc.
`Ex. 1028, p. 363
`
`

`

`364
`
`H.J. Hermens et al. / Journal of Electromyography and Kinesiology 10 (2000) 361–374
`
`and Ag/AgCl electrodes since in the presence of an elec-
`trical potential an AgCl electrode immediately becomes
`an Ag/AgCl electrode. For array or line electrodes Ag
`or Au was used.
`
`3.1.5. Electrode shape and size
`Initially, in 88 (61%) of the scanned publications the
`shape of the electrodes used was not mentioned. After
`feedback from the authors this amount was reduced to
`69 (48%). In the 75 publications (52%) the following
`shapes were used:
`
`O circular reported in 59 publications
`O rectangular/bar reported in 13 publications
`O square reported in 2 publications
`O oval reported in 1 publication
`
`Thus, in the literature both rectangular (bars) and circu-
`lar electrodes are being used for SEMG recordings of
`which circular electrode are by far the most used.
`When discussing electrode size, we have to discrimi-
`nate between circular and rectangular/bar electrodes. In
`52 (88%) of the 59 scanned publications in which circu-
`lar electrodes were reported, the size of the electrodes
`was mentioned. In some papers several electrode sizes
`were used which contributes to the fact that, in total, 57
`sizes were found. Fig. 2 shows the occurrence of the
`different electrode diameters that were found.
`From Fig. 2 it becomes clear that there is a slight pref-
`erence to use circular electrodes with a diameter ranging
`from 8 to 10 mm. It can also be concluded that an almost
`continuous range of electrode diameters was used. Only
`two occurrences relate to array electrodes (1 mm diam-
`eter (once), 2 mm diameter (once)). All other occur-
`rences relate to the electrodes used in mono- or
`bipolar recordings.
`With respect to the square/rectangular/bar electrodes
`it is not possible to detect any particular preference. The
`size of 12 of the 15 electrodes found in the literature
`
`was mentioned, usually expressed as width (mm)· length
`(mm). Although the orientation of an electrode with
`respect to the fibers is of importance (which side—long-
`est or shortest side—is placed perpendicular to the mus-
`cle fibers), this was badly described. As such, the defi-
`nition of ‘width’ in this chapter means nothing else but
`the shortest length of the electrode. The following over-
`view shows that a large variety in sizes was used.
`
`array)
`array, 2· bipolar)
`array)
`
`O 1· 5 mm (1·
`O 1· 10 mm (1·
`O 2· 10 mm (1·
`O 3· 5 mm (1· )
`O 4· 7 mm (1· )
`O 5· 10 mm (1· )
`O 5· 12 mm (1·
`O 6· 12 mm (1· )
`O 11· 11 mm (1· )
`O 20· 40 mm (1· )
`
`array)
`
`3.1.6. Inter-electrode distance
`The effect of the inter-electrode distance (IED) on
`SEMG signal characteristics is regarded as one of the
`most relevant property of the SEMG sensor. Fig. 3
`shows an overview of the different IED occurrences
`found in the scanned publications, both for line/array and
`for bipolar electrode configurations.
`A high variability and a wide range of values for IEDs
`were found. One could expect
`that
`larger distances
`would be used for larger muscles. This seems not to be
`true; for most of the larger muscles the whole range of
`electrode distances was found (i.e. biceps brachii 10–40
`mm, biceps femoris 20–50 mm, deltoideus 20–40 mm,
`gastrocnemius 10–50 mm, rectus femoris 10–50 mm).
`Authors seem to have a preference for IED values which
`are a multiple of 10 mm. The largely preferred distance
`was 20 mm.
`
`Fig. 2. Occurrence of the diameter of circular SEMG electrode sizes
`found in the literature.
`
`Fig. 3. Occurrence of inter-electrode distances (in mm) of bipolar
`and array SEMG electrodes found in the literature.
`
`Petitioner - Avation Medical, Inc.
`Ex. 1028, p. 364
`
`

`

`H.J. Hermens et al. / Journal of Electromyography and Kinesiology 10 (2000) 361–374
`
`365
`
`3.1.7. Skin preparation
`In 89 (62%) of the publications the skin preparation
`technique used was not mentioned. After feedback from
`authors this amount was reduced to 68 (47%). In the
`remaining 76 (53%) publications, standard skin prep-
`aration techniques [2] were mentioned such as shaving,
`rubbing/abrasion and cleaning of the skin, or a combi-
`nation of these techniques. Rubbing/abrasion of the skin
`was done with sandpaper, glasspaper, alcohol/ether until
`redness. Cleaning was done using alcohol, ethanol, ether,
`acetone, a mixture of these products or a cleaning gel.
`A total of 18 publications (13%) indicated that the skin
`impedance was checked before SEMG recordings were
`taken. Fig. 4 shows an overview of the maximum imped-
`ance, which was accepted in those publications. Apart
`from one exception (100 kV) all authors accepted a
`maximum skin impedance below 10 kV.
`In 115 publications (80%) it was not clearly indicated
`whether gel was used or not. In 19 (13%) cases the
`author indicated that gel was used while the skin prep-
`aration techniques used were further detailed. In the
`remaining 10 (7%) publications it was clearly indicated
`that no gel was used at all.
`
`3.1.8. Sensor location and orientation on the muscle
`The publications were also scanned on the location
`and orientation of the bipolar sensor on the investigated
`muscle(s). In this context Location is defined as the pos-
`ition of the sensor on the muscle. It has been assumed
`that the location description described the location of the
`geometrical center of the sensor, unless specified other-
`wise. Orientation is defined as the direction of the
`bipolar sensor with respect to the direction of the mus-
`cle fibers.
`In the 144 papers scanned, in total 352 descriptions
`of the sensor location were counted. These descriptions
`applied to 53 different muscles. In four of the 352 (1%)
`descriptions neither the muscle(s) of which the SEMG
`has been recorded nor the sensor location was men-
`tioned. In 58 (16%) descriptions the sensor location on
`the muscle was not mentioned. The remaining 294
`descriptions mentioned both the name of the muscle and
`
`Fig. 4. Occurrence of the maximum accepted skin impedance of
`SEMG electrodes (in kV) after skin preparation, found in the literature.
`
`described the sensor location or referred to the literature.
`The main literature references which were found are
`[1,2,24,27,28]. These publications
`contain detailed
`sensor
`location descriptions for a large number of
`muscles. Some of the scanned publications also con-
`tained detailed sensor location descriptions for a large
`quantity of muscles [12,26].
`Tables 2 and 3 show some examples of sensor
`location descriptions for biceps brachii and gastro-
`cnemius muscles. In SENIAM 5 additional tables can be
`found for the soleus and trapezius muscles as well as
`the references to the papers in these tables.
`Table 2 shows that, in total, 21 sensor placement
`descriptions for the biceps brachii muscle were found.
`In three publications the sensor location was not men-
`tioned at all.
`Globally, three placement strategies can be discerned:
`
`1. on the center or on the most prominent bulge of the
`muscle belly (10 out of 21);
`
`Table 2
`Overview of electrode location descriptions on the biceps brachii mus-
`cle
`
`Electrode location
`
`Author
`
`Woensel W. van
`Martin A.
`Martin A.
`Clarijs JP.
`Kluth K.
`Maton B.
`
`Christensen H.
`Stegeman D.
`
`Middle of muscle belly
`To the muscle belly
`Over the belly
`Midpoint to contracted muscle belly
`?
`One of the two recording electrodes placed
`above the motor point
`Most bulky part of the long head
`Over the muscle in line with the main fiber
`direction of the short head of the muscle,
`distal to the motor point. Location accepted if
`maximum cross-correlation coefficient between
`bipolar recorded EMG signals (see below)
`.0.7
`Parallel to fiber orientation, halfway between Van der Hoeven H.
`innervation zone and distal tendon
`Fellows SJ.
`?
`Between endplate region and tendon insertion Rau G.
`On short head parallel to the muscle fibers
`Vogiatzis I.
`Belly-tendon montage
`Logullo F.
`Belly of the muscle
`Esposito F.
`Parallel to fiber orientation, halfway between Van der Hoeven H.
`innervation zone and distal tendon
`Happee R.
`In the midst of the muscle belly
`Orizio C.
`Over the belly
`Halfway between the motor endplate zone and Hermens HJ.
`the distal tendon aligned in the direction of the
`muscle fibers
`On the muscle after having determined the
`motor endplates with an electrostimulation
`apparatus
`Over the medial belly of each head (long head Perot C.
`and short head) parallel to muscle fibers
`?
`
`Kahn JF.
`
`Hummelsheim H.
`
`Petitioner - Avation Medical, Inc.
`Ex. 1028, p. 365
`
`

`

`366
`
`H.J. Hermens et al. / Journal of Electromyography and Kinesiology 10 (2000) 361–374
`
`Table 3
`Overview of electrode location descriptions on the gastrocnemius mus-
`cle
`
`Electrode location
`
`Author
`
`not mention the electrode location at all. Most authors
`clearly indicated whether the EMG of the lateral or
`medial gastrocnemius was measured and placed the elec-
`trodes on the muscle belly. Most authors described the
`sensor location used in a rather global manner: on the
`belly of the muscle, on the most prominent bulge/calf of
`the muscle, on the top of the muscle. The orientation of
`electrodes on the muscles was seldom mentioned.
`With the soleus muscle, 22 sensor placement descrip-
`tions were found. In principle, the description for this
`muscle is more critical compared to other muscles since
`the largest part of the muscle is covered by the gastro-
`cnemius muscle, so only a small part can directly be
`accessed from the surface. The literature scan showed
`that there was a large variety of locations used and it was
`difficult to categorize the electrode location descriptions.
`From a large number of descriptions it was unclear how
`to reproduce the exact electrode location (i.e. it is diffi-
`cult to determine the exact location of the muscle belly
`if most of the belly is covered by the gastrocnemius
`muscle).
`the sensor placement
`to this muscle,
`In contrast
`descriptions for the trapezius muscle were much clearer.
`With the trapezius muscle 35 sensor placement descrip-
`tions were found. In only one case was the electrode
`location not mentioned at all. In most of the descriptions
`it was clearly indicated whether electrodes were placed
`on the trapezius ascendens, transversalis or descendens
`muscle. The trapezius descendens muscle seems to be
`the most commonly investigated part of the trapezius
`muscle. In comparison to other muscles, the placement
`was referring much more often to bony landmarks, facil-
`itating a good reproduction. For the pars descendens a
`preference for a position midway between C7 and the
`acromion can be recognized.
`Hence, in general the description of the placement of
`the SEMG sensor is not good enough to enable a good
`replication of the experiment, although there are con-
`siderable differences between the different muscles. One
`could get the impression that
`in the more ‘difficult’
`muscles the description of the placement
`is worse.
`Another general comment
`is that
`the orientation of
`sensor was seldom mentioned.
`
`3.1.9. Fixation on the skin
`The way the sensor is connected to the body is
`referred to as ‘fixation’. This facilitates a good and stable
`electrode–skin contact, a limited risk of movement of
`the sensor over the skin as well as a minimum risk of
`pulling of cables.
`In general, authors hardly ever reported the fixation
`methods used. Reported fixation methods mentioned
`include the use of (double-sided) adhesive tape or collar,
`stickers, elastic bands, and keeping the sensor on the
`desired location by hand.
`
`Hainaut K.
`
`Svantesson U.
`Avela J.
`
`Voigt M.
`
`Over the motor point of the medial muscle
`(location using electrical stimulation); if
`muscle has multiple motor points, the motor
`point with the lowest threshold was chosen
`In the belly of the medial muscle group
`Longitudinally on the muscle belly of the
`lateral head
`On midpart of lateral muscle bellies,
`approximately parallel to muscle fibers
`Upper third of the leg, over the medial muscle Abbruzzese M.
`belly
`Over the most prominent bulges of the medial Peeters M.
`muscle
`Upper third of the leg, over the lateral muscle Abbruzzese M.
`belly
`Rissanen S.
`Over the belly of the muscle
`Trenkwalder C.
`On top of the muscle
`Trenkwalder C.
`?
`Portero P.
`Center of the belly of the muscle
`Nicol C.
`Longitudinally on the lateral muscle belly
`Kyrolainen H.
`Longitudinally over the muscle belly
`Separated measurement—a pair of electrodes Hof A.
`on the midst of each muscle belly, across the
`fiber direction; combined measurement—each
`electrode of the pair was on a muscle belly
`On the lateral muscle belly in the direction of Ament W.
`the muscle fibers
`Proximal electrode placed above bulkiest part
`or middle of the medial muscle belly
`Over the motor point of the lateral muscle
`(location using electrical stimulation); if
`muscle has multiple motor points the motor
`point with the lowest threshold was chosen
`On the belly of the lateral muscle group
`Longitudinally over the muscle belly;
`longitudinal distance between electrode pairs at
`least 10 cm
`Gantchev N.
`On medial gastrocnemius
`On the muscle belly of gastrocnemius medial Ament W.
`in the direction of the muscle fibers
`Over the area of greatest muscle bulk on the
`medial calf
`
`de Looze MP.
`
`Hainaut K.
`
`Svantesson U.
`Kyrolainen H.
`
`Sinkjaer T.
`
`2. somewhere between the innervation zone and the dis-
`tal tendon (6 out of 21);
`3. on the motor point (1 out of 21).
`
`In the remaining four publications the sensor location
`was not mentioned or was unclear.
`Altogether this shows that half of the authors use a
`belly-montage. Certain authors differentiated between
`the long head and the short head. The orientation of the
`electrodes with respect to the direction of muscle fibers
`was seldom mentioned.
`Table 3 shows that in the gastrocnemius muscle, the
`sensor placement is in general much clearer. In total, 22
`descriptions were found of which only one author did
`
`Petitioner - Avation Medical, Inc.
`Ex. 1028, p. 366
`
`

`

`H.J. Hermens et al. / Journal of Electromyography and Kinesiology 10 (2000) 361–374
`
`367
`
`3.1.10. Location of reference electrode
`In general, the location of the reference electrode was
`not well reported. In only 54 of the 144 (38%) publi-
`cations was the location of the reference electrode men-
`tioned. This means that for 237 of the 352 (67%) elec-
`trode location descriptions the location of the reference
`electrode is not known.
`In principle, the reference electrode is placed over
`inactive tissue (tendons or bony parts), often at a dis-
`tance from the active muscles. In some cases the refer-
`ence electrode was placed as close as possible on a bony
`part, the tendon or next to the SEMG electrodes.
`‘Popular’ locations to place the reference electrode
`were the wrist, waist, tibia, sternum and the processus
`spinosus (proc. spin.) of C7 vertebra. The wrist was
`often used while measuring the EMG of leg, arm, back,
`shoulder/neck as well as facial muscles. The waist was
`used when many shoulder/neck and back muscles were
`measured simultaneously. The tibia was used in combi-
`nation with both lower and upper leg muscles and the
`sternum in combination with back muscles and the pro-
`cessus spinosus (proc. spin.) of C7 vertebra in combi-
`nation with shoulder/neck muscles.
`
`4. The effect of sensor properties and sensor
`placement procedures on SEMG characteristics
`
`During the topical workshop extensive discussions
`took place on all relevant aspects of sensors and sensor
`placement procedures with respect to their effects on
`SEMG signal characteristics. The main conclusions will
`be described in this section and the resulting recommen-
`dations will be described in the last section. A restriction
`is that this concerns only the bipolar electrode configur-
`ation. During the workshop it was concluded that elec-
`trode arrays have great potential [7,8], especially in neur-
`ology, but it is too early to develop guidelines for them.
`
`4.1. Sensor properties
`
`4.1.1. Electrode shape and size
`Although the inventory showed that circular elec-
`trodes are by far the most used, the state of the art
`showed that when considering differences only in shape
`(i.e. comparing a circular electrode with diameter R with
`a square electrode size R· R), not much difference can
`be expected. As long as the total surface area for both
`electrodes is similar the skin impedance and noise will
`also be similar.
`In conclusion there are no clear and objective criteria
`for
`recommendations for
`the electrode shape.
`It
`is
`important that the shape and size of the electrodes are
`clearly reported.
`With respect to the size of SEMG electrodes, it is
`obvious that on increasing the size, perpendicular to the
`
`muscle fibers, the impedance will decrease and it is
`expected that the view of the electrodes increases, but
`no quantitative data on the extent of this latter effect
`could be found. With respect to an increase of the size
`in the direction of the muscle fibers, it can be shown
`that this has an integrative effect on the SEMG signal,
`decreasing the high-frequency content. The effect of an
`increase of this size on the amplitude of the SEMG sig-
`nal is not clear.
`
`4.1.2. Electrode material
`Electrodes must provide good electrode–skin contact,
`low electrode–skin impedance, low noise and ‘stable’
`behavior (that is with respect to impedance and chemical
`reactions at the skin interface). The inventory has shown
`that Ag/AgCl electrodes are most commonly used. In the
`many years that these have been applied, it has become
`clear that they provide a stable transition with relatively
`low noise and are commercially available.
`
`4.1.3. Inter-electrode distance
`The signal detected by an electrode pair is the differ-
`ence between two monopolar signals and is strongly
`affected by their delay. Blok and Stegeman [5] carried
`out an interesting simulation study in which they varied
`the IED and looked at the effect on the action potential
`of a motor unit at some distance. They quantified this
`by using the area under the rectified action potential A.
`The results clearly predict that different phases can be
`discerned. Upon an increase of the IED, A will increase
`from zero to a maximum and then decrease to a plateau
`at which the action potentials are completely separated.
`This maximum will occur for an IED at which the nega-
`tive peak of one monopolar signal coincides with the
`positive peak in the other monopolar. The simulations
`indicate that this peak will occur when the IED is about
`20 mm (coinciding with about half the duration of the
`action potentials).
`With respect to obtaining a large SEMG signal and
`consequently a low signal-to-noise ratio, it seems opti-
`mal to choose the IED such that it is near this peak. An
`important condition, however, is that the electrodes will
`not be above innervation or tendon zones as this will
`lower A to another plateau value, about half that of the
`former value [5]. The IED should always be chosen such
`that the two electrodes do not approach these areas.
`The effect of the distance between motor unit and rec-
`ording electrodes has often been described in terms of
`the following function:
`V5 V0
`(r/r0)D
`where D and V0 are constants and r0 some reference dis-
`tance from the electrical center of the motor unit, where
`V=V0. This electrical center locates one equivalent gen-
`erator of the MUAPS. This equation was found by Buch-
`
`(1)
`
`Petitioner - Avation Medical, Inc.
`Ex. 1028, p. 367
`
`

`

`368
`
`H.J. Hermens et al. / Journal of Electromyography and Kinesiology 10 (2000) 361–374
`
`thal et al. [6] and Gydikov et al. [11], by fitting experi-
`mental data obtained with surface electrodes. The
`exponent D is a function of the IED. In the literature it
`has often been suggested that a decrease of the IED
`would limit the view of the surface electrodes and conse-
`quently would help to limit crosstalk. This would imply
`higher values of D upon decreasing IED. In the literature
`no evidence can be found for this. Hermens [13] col-
`lected action potentials using a spike triggered averaging
`method at two different IEDs of 20 and 40 mm. He cal-
`culated values for D but did not find any significant dif-
`ferences. More recently,