Electromyographic activity in deadlift exercise and its variants. A systematic review

The main purpose of this review was to systematically analyze the literature concerning studies which have investigated muscle activation when performing the Deadlift exercise and its variants. This study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Statement (PRISMA). Original studies from inception until March 2019 were sourced from four electronic databases including PubMed, OVID, Scopus and Web of Science. Inclusion criteria were as follows: (a) a cross-sectional or longitudinal study design; (b) evaluation of neuromuscular activation during Deadlift exercise or variants; (c) inclusion of healthy and trained participants, with no injury issues at least for six months before measurements; and (d) analyzed “sEMG amplitude”, “muscle activation” or “muscular activity” with surface electromyography (sEMG) devices. Major findings indicate that the biceps femoris is the most studied muscle, followed by gluteus maximus, vastus lateralis and erector spinae. Erector spinae and quadriceps muscles reported greater activation than gluteus maximus and biceps femoris muscles during Deadlift exercise and its variants. However, the Romanian Deadlift is associated with lower activation for erector spinae than for biceps femoris and semitendinosus. Deadlift also showed greater activation of the quadriceps muscles than the gluteus maximus and hamstring muscles. In general, semitendinosus muscle activation predominates over that of biceps femoris within hamstring muscles complex. In conclusion 1) Biceps femoris is the most evaluated muscle, followed by gluteus maximus, vastus lateralis and erector spinae during Deadlift exercises; 2) Erector spinae and quadriceps muscles are more activated than gluteus maximus and biceps femoris muscles within Deadlift exercises; 3) Within the hamstring muscles complex, semitendinosus elicits slightly greater muscle activation than biceps femoris during Deadlift exercises; and 4) A unified criterion upon methodology is necessary in order to report reliable outcomes when using surface electromyography recordings.


Introduction
Resistance training provides several health benefits related to enhancing muscle strength, reversing muscle loss, reducing body fat, improving cardiovascular health, enhancing mental health and increasing bone mineral density [1][2][3][4][5][6]. Accordingly, resistance training should be considered essential for the whole population, but it is even more relevant when the target is the transference into some specific activity or daily life tasks [7,8], injury prevention [9] or maximizing sports performance [10]. Free-weight resistance training is already well known as a key point in every strength training program [11][12][13]. In categories of creating diverse stimulus for muscle groups, different modalities such as barbell, kettlebells, hexagonal bars or dumbbells devices are typical recurring resources for coaches and trainers [14,15]. Besides, other implements which can considerably modify the exercise load profile are elastic bands [2,25], chains [29] or Fat Gripz devices [24].
It is essential to be acquainted with which muscles are activated during certain exercises and to compare different movement patterns when choosing exercises for a concrete objective [16]. Surface electromyography (sEMG) is one of the main tools used to measure muscle activation, and it can be defined as an electrophysiological recording technology used for the detection of the electric potential crossing muscle fiber membranes [17]. Thereby, task-specific data regarding motor unit recruitment patterns are reported through sEMG. For instances, athletes have the possibility to perform a concrete exercise when targeting a particular muscle [18,19].
Deadlift, Squat and Bench Press are basic resistance exercises performed in several training programs for improving physical fitness in athletes [20]. This explains the great interest in studying muscle activation, which also translates these movements into some of the most investigated exercises in the current literature using sEMG [14,21,22]. Deadlift is frequently performed primarily when the goal is the strengthening of thigh and posterior chain muscles; specifically gluteus, hamstrings, erector spinae and quadriceps [23,24]. Thus, Deadlift is classified as one of the most typical resistance exercise for posterior lower limb strengthening, as well as its variants [25]. Moreover, Deadlift has been mentioned in numerous studies comparing this exercise with other variants such as Stiff Leg Deadlift [26], Hexagonal Bar Deadlift [22] or Romanian Deadlift [27]. It has also been contrasted with other less popular variants such as Sumo Deadlift [13], unstable devices [28] and elastic bands Deadlift [8], among others.
To the best of our knowledge, there is no comprehensive review of the current literature concerning Deadlift movement pattern, and there is significant controversy when determining which muscles are involved within each Deadlift variants. For instance, the greatest muscle activation has been reported for the biceps femoris compared with the erector spinae and gluteus maximus during Deadlift [8], whereas Snyder et al. (2017) found greater erector spinae activation in comparison with gluteus maximus and biceps femoris. In contrast, Andersen et al. (2018) reported maximal activation for biceps femoris versus gluteus maximus and erector spinae for the same tested movement.
Thus, the main purpose of this manuscript was to systematically review the current literature investigating muscle activation measured with sEMG of muscles recruited when performing the Deadlift exercise and all its best-known variants. An increased understanding of the muscle activation that occur during these exercises will provide the researcher, clinician and athletes with relevant information about the use of the best exercise to activate a specific muscle or group of muscles associated with the Deadlift and its variants.

Methods
This systematic review was reported and developed following the Preferred Reporting of Systematic Reviews and Meta-Analysis (PRISMA) guidelines [29,30]. The protocol for this systematic review was registered on PROSPERO (CRD42019138026) and is available in full on the National Institute for Health Research (https://www.crd.york.ac.uk/prospero/display_ record.php?ID=CRD42019138026). The quality of included studies was assessed by two reviewers using the PEDro quality scale, which consists on eleven questions and distributes the score proportionally to the total amount of questions included. However, due to the inability to blind researchers and trainees, three of eleven questions were excluded from the scale resulting in a maximum of eight [17].
A literature search of PubMed, OVID, Scopus & Web of Science electronic databases was performed from March-April 2019. Reviews included publications from inception until March 2019.
Studies were included if they met the following criteria: i. cross-sectional or longitudinal (experimental or cohorts) study design; ii. evaluated neuromuscular activation during Deadlift exercise or variants; iii. included healthy and trained participants, with no injuries for at least six months before measurements; iv. analyzed "sEMG amplitude", "muscle activation" or "muscular activity" with surface electromyography devices (sEMG); Most articles found were written in English, but there were no language restrictions. Reviews, congress publications, theses, books, books chapters, abstracts, and studies with poor protocol description or insufficient data were not included. Studies whose participants did not have at least six months of resistance training experience were excluded. We also excluded all studies in which participants were under eighteen years old due to underdevelopment of strength and coordination [31]. Studies reporting muscle activation only from upper limbs during Deadlift exercise were also considered.
As different terms are related to the same concept, in categories of unifying criteria, the "muscle activation" term will be used when referring to "sEMG amplitude", "muscle excitation", "muscle activity", "neuromuscular activity" or similar.
Articles were selected by two independent reviewers according to inclusion and exclusion criteria. After eliminating duplicates, the titles and abstracts were analyzed and if there was not enough information, the full text was evaluated. All studies identified from the database searches were downloaded into the software EndNote version X9 (Clarivate Analytics, New York, NY, USA).
Every decision was approved by both reviewers. However, a third reviewer was consulted in case of disagreement. The whole search process took two weeks. All steps taken are thoroughly described in the flow chart (Fig 1).
Muscle activation was the main data gathered, dividing eccentric and concentric sEMG activity data when reported. All studies finally selected reported muscle activation of every muscle and exercise separately. Furthermore, data related to exercise loading and exercise description details were collected.
Data collected in this review could not be analyzed as a meta-analysis since there was not enough homogeneity in terms of the type of analysis and methods carried out amongst studies. Therefore, a qualitative review of the results was conducted.

Search results
A total of 207 articles were identified from an initial survey executed by two independent reviewers. 98 of these articles were duplicated, which led to a remaining amount of 109 in the process. The next step involved reading the title and abstract with the purpose of eliminating all those not meeting the inclusion criteria. Finally, twenty-eight articles were fully read, and nineteen of these were eventually selected for the review (Fig 1). The publication date of all selected articles ranged from 2002 to January 2019. Additionally, all studies were categorized as having a good/excellent quality in the methodological process based on the PEDro quality scale.
All selected articles presented a cross-sectional design. In fact, most experimental studies found used an untrained participant sample, so they were excluded. Regarding experience time, all participants had at least six months of previous resistance training experience, although some studies did not report the exact experience time of participants (Table 1).
No common criteria were followed when referring to the exercise loading at which exercises were evaluated during sEMG recordings. As a matter of fact, only two studies used a similar method, assessing one repetition maximum intensity (1RM) [22,32]. Some studies measured a number of repetitions of xRM, whereas others measured a number of repetitions of a range between 65-85% of 1RM (Table 1), which could be considered in all cases as a submaximal load intensity [33].
Data regarding the studies' general description and main findings are presented in Table 1, while Tables 2-5 contain data referring to muscle activation during Deadlift exercise and/or its variants. We found no unified criteria for the sEMG normalization method. Out of all included studies, seven reported data description regarding muscle activation in relation to exercise type and normalized sEMG activity as a percentage of maximal voluntary isometric contraction (% MVIC) ( Table 2); three of them as percentage of peak root mean square (% peak RMS) ( Table 3); two studies reported data expressed as absolute RMS values in microvolts (mV) ( Table 4); and three studies expressed data as a percentage of 1 repetition maximum (% 1RM) ( Table 5). In addition, there were four studies which were not included in the   [16,43], Romanian Deadlift (2/19 studies) [27,44] and Hexagonal Bar Deadlift (2/19 studies) [22,38] (Table 1).
It is also important to clarify that exercises such as "Olympic Barbell Deadlift", "Straight Bar Deadlift", "Barbell Deadlift", "No Chains Deadlift" and "Conventional Barbell Deadlift" all refer to the same exercise, so "Deadlift" will be used for all cases.

Concentric and eccentric phases
Generally, studies analyzing electromyographical data assess muscle activation on each repetition, treating it as a single unit. Nonetheless, it has been reported that electromyographical activity could differ significantly between concentric and eccentric phases of the movement. Therefore, some authors have already carried out this division in their research [45][46][47]. Not all studies included in the current review divided sEMG exercises into concentric and eccentric phases. In fact, only seven studies performed such a subdivision [16,28,34,37,38,40,41], in which the concentric phase showed greater muscle activation than the eccentric phase for every single case.

Muscle activation
The biceps femoris has been the most investigated muscle in terms of sEMG for the Deadlift exercise and its variants (13/19 studies). Gluteus maximus is the next muscle most evaluated n/a n/a 39.8 ± 16.6% mean conc 19.9 ± 11.3% mean eccen n/a n/a n/a n/a Andersen et al. (2018) [22] Deadlift~95% mean~108% mean n/a n/a n/a n/a n/a~86% mean Hexagonal bar deadlift (HBDL)~8 8% mean~83% mean n/a n/a n/a n/a n/a~82% mean Iversen et al.

Schoenfeld et al. (2015) [42]
Stiff leg deadlift n/a~40% mean lower~73% mean upper~4 7% mean lower 125% mean upper n/a n/a n/a n/a n/a Ebben (2009) [43] Stiff leg deadlift n/a n/a n/a 49±27% mean n/a n/a n/a n/a Unilateral stiff leg deadlift n/a n/a n/a 48±39% mean n/a n/a n/a n/a Good morning n/a n/a n/a 43±16% mean n/a n/a n/a n/a  (Table 1). Due to the diversity regarding methodology, it was considered appropriate to report the results by grouping the studies according to the sEMG normalization process carried out in each study (mean or peak % MVIC, % peak RMS, RMS mV or % 1RM).
Studies in which muscle activation was expressed as a mean or peak % MVIC are shown in Table 2. Erector spinae showed the greatest muscle activation during the Stiff Leg Deadlift exercise [41], and also showed a similar muscle activation than the gluteus maximus or biceps femoris during Deadlift and Hexagonal Bar Deadlift exercises [22]. Except for the Deadlift exercise [22], the gluteus maximus showed greater muscle activation than biceps femoris [13, [16,41,42], which is even more pronounced when performing Unilateral Stiff Leg Deadlift [16]. The concentric phase showed a greater activation in the gluteus maximus and hamstring muscles than the eccentric phase for all exercises evaluated [16,40,41] (Table 2). Data regarding muscle activation expressed as percentage peak RMS (% peak RMS) are shown in Table 3. The erector spinae and lumbar multifidus showed greater muscle activation than the gluteus maximus and biceps femoris [26,36]. However, conflicting results have been reported for the Deadlift exercise. Lee et al. [27] reported more activation in the biceps femoris than the gluteus maximus, while Snyder et al. [36] reported more activation in the gluteus maximus than the biceps femoris (Table 3). Whereas the vastus lateralis showed greater muscle activation than the biceps femoris [26,36], and the rectus femoris showed greater muscle activation than the biceps femoris and gluteus maximus during Deadlift exercise [27] (Table 3).
Data regarding muscle activation expressed as RMS in mV are shown in Table 4. Erector spinae and semitendinosus are the most activated muscle in the Deadlift exercise [22]. When comparing muscle activation within the hamstrings, there was a greater activation recorded for the semitendinosus muscle in comparison to that for the biceps femoris [22,44] (Table 4). The concentric phase showed greater activation than the eccentric phase in all muscles and exercises evaluated [44].
Data regarding muscle activation in mV expressed as a percentage of sEMG (mV) during a 1RM effort are shown in Table 5. The Erector spinae presented higher muscle activation than the gluteus maximus [37] and biceps femoris [38]. The vastus lateralis and vastus medialis showed greater muscle activation than the biceps femoris and gluteus maximus during Deadlift exercises and its variants [37][38][39]. The concentric phase showed greater activation in the biceps femoris, vastus lateralis and erector spinae than the eccentric phase during the Deadlift exercise as well as during the hexagonal bar Deadlift exercise [38].
Erector spinae presented higher muscle activation than the gluteus maximus and the biceps femoris muscles for all exercises [8,26,37,38,41]. Only one study presented contrary outcomes, showing lower muscle activation in the erector spinae than the biceps femoris and semitendinosus during the Romanian Deadlift exercise [36].
Another important finding in the current review was that muscles from the quadriceps complex appeared to elicit the greatest muscle activation compared to the gluteus maximus and hamstrings muscles for Deadlift exercise [26,27,[36][37][38][39]. Furthermore, the semitendinosus generally tended to elicit slightly greater muscle activation than the biceps femoris within the hamstring complex [8,44].

Methodological issues
One concern about the findings of the review was the lack of unification of collecting data methodological process amongst studies. This includes the kind of muscle contraction evaluated, the number of participants, the participants' resistance training experience, exercise intensity during evaluation, sEMG collection method, electrode location, and number of evaluation days. All studies following a specific methodology process had diverse aims and different outcomes, which made difficult to deliver consistent results. Only one study evaluated just an isometric position of the movement, the preparatory position [35], whereas the rest evaluated exercises from a dynamic perspective.
In line with previous reviews, exercise loading for sEMG recordings has been one of the biggest concerns [52][53][54]. Only two studies performed same 1RM intensity [14,25], whereas others performed exercises at a predetermined repetition maximum load, and the rest measured a number of repetitions within a range of 65-85% 1RM (Table 1). Differences in the applied methodology should be reduced for future studies, providing an enhanced outcomes reliability [42].
No unified criterion has been followed in categories of time management during exercise phase among study methodologies, which could also be treated as a potential bias risk. For future studies focused on sEMG, it would be of significant interest to report divided electromyographical data into concentric and eccentric phases, as well as exercise timing. Such information would help coaches and trainers when choosing one or another exercise for a concrete target when prescribing an optimized training [7].
In relation to the electrode location, reports on surface recording of sEMG should include electrode shape and size, interelectrode distance, electrode location and orientation over muscle with respect to tendons and fiber direction among others (Merletti & Di Torino, 1999). It is vital to report in detail the placement of electrodes over the muscle belly when we aim to compare outcomes with other similar studies.
Different protocols for surface electromyography electrode placement have been described in the literature. One of the most popular protocols is the SENIAM Guidelines (Surface ElectroMyoGraphy for the Non-Invasive Assessment of Muscles). Eight studies following the SENIAM Guidelines have been included in our review [8, 16, 22, 35-37, 41, 42]. The rest followed some other Guidelines or a previous reference, and only four studies did not report any protocol for electrode location [26,28,32,34].
In regard to interelectrode distance, five studies reported using the recommended 2 cm center-to-center distance between electrodes according to SENIAM Guidelines [8,22,27,35,43]. In addition to not following these Guidelines, some other studies also did not report the inter-electrode distance [26,32,34,[38][39][40]44]. Furthermore, four studies reported to have placed the electrodes with a center-to-center distance ranging between 15-35mm but different from 20 mm [13,16,28,37,41]. The higher the interelectrode distance, the wider the detection volume and consequently the detected amplitude [55]. Future research should attempt to follow established Guidelines, so they can reach optimum research quality and diminish the risk of data collection bias.
On the other hand, most of the reviewed studies included between 2-4 days/sessions (visits to the laboratory) for the measurement process, normally leaving 2-7 days' rest between each visit. Tasks performed during those days cover anthropometric data gathering, familiarization with exercises, RM testing and sEMG data collection. To ensure reliable sEMG data outcomes, sEMG data must be collected at the same session [8]. Otherwise, some studies collected sEMG data on two different days, which might have entailed electrode location mistakes [32,40,41,44].
In order to avoid fatigue bias risks, a randomized counterbalanced order for exercise testing was followed in all studies but one, which followed a preset exercise order [41]. In addition with the same aim, a minimum break of 2-5 min was considered between exercise testing trials [22,27,28,34,36,37,[40][41][42]44].
Most studies did not report hand grip and stance position in any depth of detail. Some studies allowed a preferred stance position for each participant but maintained the same for all exercises tested [8,22], whereas others also indicated a hand grip slightly wider than shoulder width [26,27,32,[40][41][42][43].

Other Deadlift variations
Apart from the above-mentioned Deadlift exercises, there are some other studies which focused on less conventional variants of this movement. The Good Morning exercise appears to be an appropriate substitute to Romanian Deadlift when it is preferable to place the load on the back instead of lifting it from the floor. Good Morning provokes a similar muscular pattern activation as Romanian Deadlift, but it showed more muscle activation for the semitendinosus and less muscle activation for the biceps femoris than Romanian Deadlift [44].
In addition, some authors proposed interesting alternatives for the Deadlift exercise with the goal of overcoming the sticking region. This involves a phase during the lift in which there is a mechanical disadvantage that elevates injury risk and leads to a deceleration on the speed lift [56]. In relation to this issue, Nijem et al. Elastic bands have also been used as a tool in Deadlift learning processes, when the athlete is not ready to lift high loads with a proper technique or in those cases when some injury prevents the athlete from using conventional resistance equipment. Muscular activation presented during elastic bands Stiff Leg Deadlift was lower than that elicited during free weights Stiff Leg Deadlift, with significant differences when referring to the gluteus maximus, biceps femoris and semitendinosus muscles [41].
Furthermore, it should be noted that if your aim is to increase muscle activation from forearm musculature during the Deadlift exercise, it is recommended to use a Fat Gripz device, a wider grip implement that sticks to the bar. Worth mentioning that a significant reduction in 1RM strength would appear when using this kind of implement [32]. Moreover, the Hip Thrust exercise has also been found to elicit greater muscle activation for the gluteus maximus than Deadlift and Hexagonal Bar Deadlift. Also, lower muscle activation for the biceps femoris muscle was shown during Hip Thrust compared to Deadlift. No muscle activation differences were presented among those three exercises for the erector spinae muscle. Hence, a greater torque and greater stress in the hip joint during Deadlift compared to both other exercises was also reported [22].

Comparing Deadlift to other exercises
Additionally, several authors have compared Deadlift exercises to single joint and machinebased exercises in their research. For example, Bourne [43].
On the other hand, the Prone Leg Curl in machine was found to elicit higher muscle activation for both upper and lower sections of the biceps femoris muscle than during Stiff Leg Deadlift but showed no significant differences for the semitendinosus muscle (Schoenfeld et al., 2015). On the contrary, McAllister et al. (2014) reported greater biceps femoris muscle activation during Romanian Deadlift than during the Prone Leg Curl. It would be necessary to unify the muscle activation normalization method and protocol carried out. Likewise, researchers should ascertain a proportional exercise load when comparing bilateral multi joint exercises to single leg and machine-based exercises, in order to obtain consistent outcomes.
Some recommendations for future research involving surface electromyography recordings are: 1. Participants training status and participants resistance training experience should be outlined in detail. Only 11/19 studies showed this information.
2. Exercise load quantification method during sEMG recordings must be standardized, so exercises could be comparable among them.
3. Taking into consideration the different muscle activation pattern reported during concentric and eccentric exercises phases, it is highly recommended to perform such subdivision for future studies.

4.
A unified criterion upon methodology protocol is necessary in order to avoid several bias risks and report reliable outcomes when using surface electromyography recordings. Information regarding electrode location, number of testing days and sEMG normalization method should be strictly reported.

Practical applications
Currently, Deadlift is an exercise frequently performed to improve the lower limb muscles, mainly biceps femoris and semitendinosus (hamstrings), and gluteus maximus. Based on this systematic review about the sEMG activity in the Deadlift exercise and its variants, it has been demonstrated that other muscles such as erector spinae and quadriceps are more activated than hamstrings and gluteus maximus, although some studies found conflicting results.
Deadlift exercise comprises a movement which could have a transference into daily life activities; also considered as one of the greatest compound lifts, as it involves several muscles groups coordination. A broad spectrum of Deadlift variants has been reported, so diverse applications for these exercises could merge, covering health, rehabilitation and performance environments.
Therefore, it must be considered that muscle activation would depend on the Deadlift variant performed. For instances, posterior thigh muscles would show greater muscle activity when performing exercises that holds the knees on a fixed and extended position (e.g. Romanian Deadlift or Straight Leg Deadlift). On the contrary, whether your goal is to maximize anterior thigh and lower back muscle activity, Deadlift would be the exercise of choice. Hexagonal Bar Deadlift also elicits a great anterior thigh muscle activity, but with a reduction on erector spinae muscle activity, turning this exercise into an appropriate Deadlift variant when athletes have lower back issues.
Hence, coaches, athletes and regular population ought to contemplate these findings when selecting the Deadlift exercise and its variants for their training programs, considering the individual training goals.