Article Images (0)
×![]()
featured_play_listSave article to folder
X
Create new folder
Full Text
PDF
Table of Contents
Abstract
Introduction
Materials And Methods
Statistical Analysis
Results
Discussion
Practical Applications
Conclusion
Acknowledgements
Disclosure Statement
Funding
ORCID
Data Availability Statement
Author Contributions
Abstract
Isometric knee-flexor testing is commonplace in soccer, yet data to inform choice of knee-flexion angle are limited. This study aimed to compare peak force production and muscle activation between two isometric knee-flexor tests. To compare peak force, 43 male soccer players (age 21.5 ± 5 years; stature 180.3 ± 6.3 cm; body mass 74.6 ± 8.9 kg) completed 3 × 5-second maximal efforts on the Nordbord device (Vald Performance) with a 0(ISO-Prone) and 30-degree (ISO-30) knee angle, respectively. To compare peak muscle activation, a further 13 trained male participants (25 ± 6 years; 178.2 ± 5.6 cm; 79.6 ± 13.2 kg) completed 3 × 5-second maximal efforts with wireless surface electromyography electrodes placed on the Gluteus Maximus, Adductor Magnus, Semitendinosus, Biceps Femoris (long and short heads) and Medial Gastrocnemius. Paired samples t-tests were used to detect differences in force output between tests and Pearson’s correlations to quantify associations. A Yuen’s modified t-test estimated the trimmed mean differences in muscle activation between tests. Higher peak forces were observed in the ISO-30, with mean differences of 56.4N (36.8 to 75.9 [95% CI]) and 52.9N (33.7 to 72.3 [95% CI]) for right (367 vs 314N) and left legs (351 vs 294N), respectively. Moderate correlations were observed in peak force for left (r = 0.55 (0.29 to 0.73 [95%CI])) and right (r = 0.64 (0.42 to 0.79 [95%CI]) legs. No differences in muscle activation were observed. Peak force varies between the ISO-Prone and ISO-30, with moderate associations between tests, indicating that the tests should not be used interchangeably. ARTICLE HISTORY Accepted 17 January 2025
ARTICLE HISTORY Accepted 17 January 2025 KEYWORDS Hamstring; EMG; ISO-30; ISOProne
Introduction
Hamstring strain injuries (HSI) represent a prevalent musculoskeletal issue in soccer (López-Valenciano et al. 2020). Ekstrand et al. (2023) highlighted an increase in HSI incidence (i.e., 6.7 and 3.9% for training and match-play, respectively) and burden (9.0 and 6.2% for training and match-play, respectively) between the 2014/15 and 2021/22 seasons in elite soccer. This follows previous work reporting a 4% increase in the incidence of HSI between 2001 and 2013 (Ekstrand et al. 2016), while an increase in prevalence of hamstring injuries from 12 to 24% between the 2001/02 and 2021/22 seasons has also been reported (Ekstrand et al. 2023). These data suggest that protecting and monitoring hamstring function should be a priority for practitioners. Hamstring strain injuries are multifactorial, but typically occur during high to maximal intensity sprinting between the late swing and early ground contact phases (Huygaerts et al. 2020), where the occurrence of an isometric contraction of the hamstring muscles has been proposed (Van Hooren and Bosch 2017, 2018). Within this phase, elongation of the tendon and near peak muscle activation occurs as the hamstrings transition between eccentric and concentric activity (i.e., decelerating knee extension and transitioning into a hip-extensor) (Woods et al. 2004; Linklater et al. 2010). The specific joint kinematics of this phase and the hamstring muscles’ capacity to resist forces, which is likely related to a players ability contract the hamstrings isometrically (Van Hooren and Bosch 2017), might provide further insight into a player’s HSI risk (Bahr and Holme 2003; Croisier et al. 2008; Rudisill et al. 2023; Bramah et al. 2024), although empirical evidence of this is lacking. Measuring the force producing capabilities of the kneeflexors (of which the hamstring muscles are the predominant contributor) can enhance the understanding of a player’s ability to produce such contractions. Specifically, assessing isometric strength of the knee-flexors at ‘longer length’ may provide valuable insight into HSI risk given the proposed higher muscle activation of the Biceps Femoris long head (Reurink et al. 2016) and suggestion that weakness at length is a risk factor for injury (Mendiguchia and Brughelli 2011). Data from such assessments might also contribute to decision making during rehabilitation, given the consistently reported reduction in isometric strength following HSI (De Vos et al. 2014; Charlton et al. 2018; Hickey et al. 2018). The practicality of assessing knee-flexor strength objectively has increased recently, with the availability of modern technology (Claudino et al. 2021; Asimakidis et al. 2024), with the ‘NordBord’ device (Opar et al. 2013) commonly used to assess eccentric and isometric knee-flexor strength (Bishop et al. 2022; Asimakidis et al. 2024). While eccentric knee-flexor strength (often assessed via the Nordic ‘Hamstring’ Exercise) testing is more prevalent than CONTACT Jonathan M. Taylor Jonathan.Taylor@tees.ac.uk School of Health and Life Sciences Teesside University, Centuria Building, Middlesbrough TS1 3BA, UK Supplemental data for this article can be accessed online at https://doi.org/10.1080/24733938.2025.2471316 © 2025 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this article has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or with their consent. isometric knee-flexor testing (Asimakidis et al. 2024), a poor relationship exists between performance (i.e., peak force application) on these tests (Moreno-Pérez et al. 2020). Given that the tests are designed to discriminate knee-flexor, and in particular hamstring weakness, it seems valuable to assess both rather than choosing either an eccentric or isometric assessment. Furthermore, regular isometric testing is appealing given the low risk of delayed onset of muscle soreness (DOMS), particularly in comparison to eccentric testing (Clarkson et al. 1986; Van Hooren and Bosch 2017; Chesterton et al. 2021, 2022). Currently, there is no consensus around the knee-flexion angle at which isometric strength should be assessed (Asimakidis et al. 2024). Assessments between 0- and 30-degree knee-flexion are commonplace given the link between ‘strength at length’ and injury (Mendiguchia and Brughelli 2011), and understanding the angle at which peak force is higher may give insight into the maximal strength of athletes. Knee-flexor strength may vary between these angles (Worrell et al. 2001), due to differences in the role of the hamstring muscles at different angles (Worrell et al. 2001). For instance, the Biceps Femoris muscle is reportedly near maximal activation at a knee-flexion angle around the late swing phase, and this is often the focus during hamstring assessment, given that it is the most frequently injured muscle in soccer (Reurink et al. 2016). Yet, it is unclear whether this muscle is closer to maximal activation at 0- or 30- degrees of knee-flexion. To date, limited work exploring the impact of the knee-flexion angles on muscle activation of the hamstring region via devices such as the Nordbord device exists. Clarity of hamstring muscle activation during isometric testing on the Nordbord device may inform practice and could have specific implications for assessing injury risk, rehabilitation, and exercise prescription (Delecluse 1997; Schuermans et al. 2014). This study compared peak force production during two isometric knee-flexor strength tests i.e., the ISO-Prone (0-degree knee angle) and ISO-30 (30-degree knee angle), in well-trained soccer players. A secondary aim was to compare the muscle activation of the Gluteus Maximus, Adductor Magnus, Semitendinosus, Biceps Femoris long and short heads, and Medial Gastrocnemius during these tests using surface electromyography (sEMG) in group of trained male participants.
Materials and methods
Part 1: comparing peak force outputs between the ISO-Prone and ISO-30 tests The study was conducted and reported in accordance with STROBE guidelines (See supplementary file for checklist). Using a cross-sectional experimental design, 43 well- trained soccer players (i.e., Playing at Semi-Professional Level (N = 11) or within an U18 English Premier League Academy (N = 32)) (age 21.5 ± 5 years; stature 180.3 ± 6.3 cm; body mass 74.6 ± 8.9 kg) with no lower extremity injury at the time of the study participated in two trials separated by 7 days. Testing occurred across the 2023–2024 soccer season either at the academic institution or at the training base of the respective participants depending on availability of the players and their squads to travel to the university. Where this was not possible, the researchers attended the training base of the respective clubs. Testing was conducted on a consistent day in the training cycle to facilitate players being in a similar condition for each trial and was a minimum of 48-hours post-match, to mitigate the risk of excessive fatigue. Ethical approval (ID 16,605) was granted from the institutional review board prior to study commencement. All testing was completed on the ‘Nordbord’ device (Vald Performance, Australia), sampling at a frequency of 50 hz and operated according to manufacturer’s guidelines. Assessment of isometric knee-flexor strength using the ‘Nordbord’ device is reported to have acceptable reliability for isometric knee-flexor peak force in athletes across a variety of knee-flexion angles (ICC 0.7–0.9; Coefficient of Variation ~ 8%) (Cuthbert et al. 2021; Luchner et al. 2021; Firmansyah et al. 2024). All players were familiar with the testing protocols and completed a standardized warm-up prior to testing, adapted from Kadlec et al. (2019). Players completed the tests in a counterbalanced order to reduce the risk of systematic bias (i.e., half of the sample completed ISO-Prone first, and half the ISO-30). Players took position on the ‘Nordbord’ device as described by Bishop et al. (2022), i.e., with their ankles fixed in place by individual hooks attached to loads cells secured immediately superior to the lateral malleolus and allowing force production to be assessed. Participants completed three 5-second maximal isometric contractions, interspersed by 30-second rest. During the ‘ISO-Prone’ test, players assumed a prone position with elbows resting on the ground and were instructed to maintain this position, with a 0-degree knee angle throughout each trial. During the ISO-30 test, players knees were flexed at a 30-degree angle throughout the test, assessed using a goniometer for each trial, with their upper body in push-up position (as described by manufacturer guidelines). The highest peak force produced during a single trial in Newtons (N) for each limb was retained for analysis. Testing positions are displayed in Figure 1. Part 2: muscle activation during the ISO-Prone and ISO-30 tests We sought to establish the muscle activation, using sEMG during the ISO-Prone and ISO-30 tests. Using a cross-sectional observational design, further 17 trained male participants (free from injury) (age 25 ± 6 years; stature 178.2 ± 5.6 cm; body mass 79.6 ± 13.2 kg) were recruited. Participants attended the laboratory at the academic institution on two occasions, for familiarization and data collection sessions, respectively. During the data collection session, Trigno Wireless System EMG sensors with silver 4-bit contact (5×1 mm dimension) were applied. Electrodes were placed on the dominant limb along the axis of the muscle fibers for the Gluteus Maximus (GM), Adductor Magnus (AM), Semitendinosus (ST), Biceps Femoris long and short heads (BFl and BFs), and Medial Gastrocnemius (MG) in accordance with the SENIAM guidelines (http://www.seniam. org/, 2024). A ground electrode was placed on the Medial Malleolus. Prior to placement the skin was shaved, cleaned with an alcohol wipe and the superficial skin cell debris removed with hypoallergenic tape. The sEMG signals were recorded by an EMG acquisition system (Delsys Trigno Wireless EMG, Delsys Incorporated, Natick, MA, USA), with a sampling rate of 2000 hz (using Delsys EMGworks Acquisition V 4.8.0 and Delsys EMGworks Analysis V 4.7.3.0 software systems; CMRR = minimum 200 mV with a 60 hz frequency, baseline noise > 750 nV) and converted from an analog to digital signal (16 bit), with full wave rectification. The digitized sEMG data were band-passfiltered at 20–450 hz using a Butterworth filter. For muscle activation time domain analysis, the root mean square (RMS) (30 ms moving window) was calculated. Participants completed three maximal trials (interspersed with 30-s rest) of the ISO-Prone and ISO-30 in a single session, in a counterbalanced order across the group. The decision to have participants complete both tests in the same session was taken to reduce the risk of error in electrode placement when re-applying on a different date. Participants were allowed 10- minute recovery between tests, to reduce the risk of potentiation influencing data (Baudry and Duchateau 2007), while allowing adequate recovery. Comparison was made between the muscle activation for each site during the ISO-Prone and ISO-30.
Statistical analysis
All data were processed through R-Studio (version: 2022.12.0 + 353); example code is provided in the supplementary materials. Part 1: A general linear model was fitted to explore the relationship between peak force during each test. The behaviour of the residuals was visually inspected and met the assumptions of normality and homogeneity of variance (see supplementary materials 2). Therefore, a t-test was performed to compare differences between ISO-Prone and ISO-30 tests, which were presented as raw effect sizes and 95% confidence intervals and exact p-values to 3 significant digits as a measure of compatibility of our observed data under the background statistical assumptions, including the null hypothesis (Greenland et al. 2016). The values within the 95% confidence interval are interpreted as highly compatible with our data under these assumptions (Greenland et al. 2016). P-values were subsequently converted into surprisal values (s-value) by taking the negative base-2 logarithm of the p-value [S-value = -log₂(p-value)] (Rafi and Greenland 2020). Surprisal values provide bits of information against the model assumptions, which include the nullhypothesis (ISO-Prone = ISO-30) and are equivalent to flips of a fair coin. For example, a surprisal of 4 suggests that the chances of observing a difference as larger or larger if the null hypothesis (and all other assumptions) is true, are no more likely than observing four heads on a fair sided coin. To estimate the level of association between peak force on the ISO-Prone and ISO-30 tests, a Pearson’s correlation coefficient was calculated and presented in standardised units with 95% confidence intervals. Descriptors were based on the following thresholds: > 0.90, very strong; 0.89–0.70, strong; 0.69–0.40, moderate; 0.39–0.10, weak (Schober et al. 2018). Prediction equations were derived from the linear model using the equation y = (x*slope) + intercept and the standard error of the prediction were reported. Part 2: Due to irregularities in the sEMG data for 4 of the participants, we removed their data from the analysis, therefore a total of 13 participants data sets were included for analysis. A general linear model was fitted to explore the relationship between sEMG during the ISO-Prone and ISO-30 tests. Here the peak root mean-squared sEMG signal was averaged across the three trials. These data (both raw and log-transformed) demonstrated visibly pronounced skewness and kurtosis, and the behaviour of the residuals did not meet the assumptions of normality and homogeneity of variance. Therefore, we used Yuen’s modified t-test for trimmed means via the ‘WRS2’ (Mair and Wilcox 2020) package and presented summary statistics as the median and interquartile range. We compared 20% trimmed means (Field and Wilcox 2017) and derived percentile (95%) bootstrap confidence intervals from the default 599 bootstrap samples.
Results
Part 1 Peak force was clearly higher during the ISO-30 than ISO-Prone test (Table 1), and while peak force in the two tests was correlated, r values deemed highly compatible with our data ranged from 0.29 (weak) to 0.72 (strong) for the left leg and 0.42 (moderate) to 0.78 (strong) for the right (Figure 2). The standard error for the prediction was large suggesting it would be difficult for practitioners to accurately predict one from the other (Table 2). Part 2 We did not observe strong evidence to refute the null hypothesis for any sEMG site. Our data provided no clear refutational information (i.e., flips on a fair coin toss) against the null hypothesis for the hamstring musculature and less than 2 bits of refutational information for other sites except for abductor magnus where the surprisal was still less than 3 (Table 3).
Discussion
This study compared the peak force production during two commonly used isometric knee-flexor strength tests (i.e., ISOProne and ISO-30) in a group of well-trained soccer players. A secondary aim was to compare the muscle activation during these tests using sEMG. Our data indicate that peak force production was significantly higher during the ISO-30 test, with ~ 50N difference between tests reported for each limb. We observed moderate correlations in peak force between the tests. Prediction equations exhibited large standard errors (~50 to 60 N) suggesting the tests cannot be used interchangeably when monitoring general isometric knee-flexor strength properties of individual players. No differences in muscle activation were observed between tests at any of the muscle sites assessed (i.e., Gluteus Maximus, Adductor Magnus, Semitendinosus, Biceps Femoris, Medial Gastrocnemius), this suggests that other mechanisms may underpin the differences in force production between tests. Assessing isometric knee-flexor strength provides potentially valuable information regarding performance, HSI risk and can act as an objective marker during rehabilitation (Van Hooren and Bosch 2017; Moreno-Pérez et al. 2020). This is illustrated by the strong correlations reported between isometric knee-flexor strength and maximal power output (r = 0.83) in professional soccer players (Boraczyński et al. 2020), and the reported 44.8N reduction in isometric knee-flexor strength in athletes with past HSI (Charlton et al. 2018). The increased prevalence of technology facilitating accurate assessment of knee-flexor strength such as the Nordbord device, has made regular testing feasible. While previous work has assessed peak force production using different testing positions (specifically knee-flexion angles), to our knowledge, this is the first study exploring the differences between the ISO-30 and ISO-Prone tests using the ‘Nordbord’ device. The greater peak force production observed (i.e., ~50 N higher) during the ISO-30 in comparison to the ISO-Prone test somewhat supports the findings of Onishi et al. (2002), who observed peak torque of the knee-flexors to occur between 21–49 degrees and Worrell et al. (2001), who suggested that peak knee-flexor torque occurs between 10–30 degrees. Yet our findings contradict work by Kumazaki et al. (2012) who reported peak torque occurred at 0 degrees of knee-flexion. Although the differences in outcome measure may limit direct study comparison, this provides some context for our data. The increased force we observed during the ISO-30 may be explained by the hamstring muscle force-length relationship (Kellis and Blazevich 2022). With respect to the active tension of the hamstrings, it is suggested that optimal active capacity occurs at a knee-flexion of ~ 30 degrees, with a hip flexion angle of 45 degrees (like the ISO-30 test) (Kellis and Blazevich 2022). It could be theorized that the ISO-30 presents a more favourable position biomechanically for knee-flexion than the ISO-Prone. The knee joint angle during the early ground contact phase of sprinting is ~ 27.7 degrees (Miyashiro et al. 2019), which provides further context regarding the enhanced force output observed during the ISO-30. The nature of soccer training and match-play whereby high-speed running and sprinting is commonplace, may provide a specific stimulus which is manifested in the forces produced during isometric testing. We observed moderate associations between peak force on the ISO-30 and ISO-Prone which suggests that strong players on one of these tests, will likely be stronger on the other. This is logical given the small difference in knee-flexion angle and the nature of the tests. However, the point estimates for these correlations suggest that the variance in one test explains only 30 and 41% (r2) of the variance in the other. Indeed, such correlations have been said to represent ‘very poor’ validity (Serpiello et al. 2017). We compared two practical measures (method comparison) and not a practical measure against a criterion or gold standard. As such, there is inevitable noise in each measure that will reduce these correlations and may contribute to unexplained variance between the tests and subsequently poor prediction interval. For instance, the typical error of 8% has been reported previously for the ISO-Prone (Luchner et al. 2021), which would represent ~ 30N in our population. It is also possible these tests assess different strength properties or involve differences is muscular contribution (Onishi et al. 2002). While both tests appear to track isometric knee-flexor strength as a general concept, practitioners are advised against using the tests interchangeably. It is accepted that the hamstrings operate disparately at varying lengths, and during different exercises (Bourne et al. 2016; Kellis and Blazevich 2022), and knee-flexion angle may affect posterior chain muscle activation during an isometric hamstring strength assessment (Worrell et al. 2001; Read et al. 2019). Specifically, Worrell et al. (2001) reported significant differences in hamstring activation between 0- and 30-degree angles, but no difference in Gluteus Maximus activation between these angles. Similarly, Kwon and Lee (2013) reported differences in hamstring muscle (i.e., Biceps Femoris and Semitendinosus) activation between knee angles, while Onishi et al. (2002) reported peak Biceps Femoris activation occurs between 15 and 30 degrees of knee-flexion. These findings contrast ours with respect to hamstring muscle activation (i.e., Biceps Femoris long and short heads, Semitendinosus). Some of the disparity may be explained by the differences in testing equipment/positions and the differences in electrode placement between the studies. It should be acknowledged that significant variability in muscle activity during isometric kneeflexion tests has been reported within participants session-tosession and side-to-side (Krishnan and Williams 2009), and it is reasonable to suggest that variation between participants may be expected (Read et al. 2019). Within-subject variability may, to some extent, be explained by the instructional cue use, with our participants instructed to ‘push’ against the cuffs as hard as possible through knee-flexion, which may be interpreted differently between participants i.e., hip extension rather than knee-flexion, and consequently limit consistency of muscle activation (Read et al. 2019). Similarly, while we ensured that knee-flexion angle was at 30 degrees during the ISO-30 test, differences in anthropometric characteristics between participants mean that the testing position cannot be highly precise (i.e., hip-flexion angle and upper body position). This may also have contributed to the variability in muscle activation, due to the changing role of the involved muscles. As indicated, previous work exploring the effects of knee angle during isometric knee-flexor strength tests on peak force/ torque and muscle activation has not used the Nordbord device, and direct comparisons should be interpreted with caution. Specifically, differences in hip-flexion angle, position of the upper body i.e., prone vs supine or upright, and the force transducer attachment point (i.e., calf or ankle attachment/cuff) may influence outcomes. However, the increasing use of the ‘Nordbord’ device for such testing procedures in soccer clubs indicates the need for research in this area, increasing the applicability of our findings. While our work provides novel findings, it is not without limitation. We acknowledge that the sample used in part 2 of the study represents well-trained participants and not welltrained soccer players, and that the sample size is limited in this part of the study. It is of note that we recruited 17 participants for part 2 of the study; however, due to irregularities in some the participants sEMG data, we chose to exclude 4 participants. Despite this, the data presented are valuable, given the similarity in muscle activation between populations reported previously during isometric testing; suggesting that translation of the findings is possible (Nimphius et al. 2019). Finally, we acknowledge that our decision to report peak force rather than peak torque deviates from typical outcome measure reporting with respect to isometric knee-flexor strength testing, although this is the correct metric to investigate our hypothesis. The ‘Nordbord’ directly measures force through a force transducer; torque is subsequently calculated from the knee position as torque is equal to force multiplied by the distance from the fulcrum (knee). This contrasts with dynamometry which directly measures torque.
Practical applications
The feasibility of objective assessment of knee-flexor strength in the field has increased, with the ‘Nordbord’ device now commonly used in practice. Our data provide valuable practical applications during the assessment of isometric knee-flexor strength using this device. Firstly, our data provide clear indication regarding the importance of accurate assessment of knee-flexion angle, with substantial variability between peak force outputs with a variation in knee-flexion angle of as little as 30 degrees. Practitioners must ensure precision with respect to the knee-flexion angle, using suitable technology to verify this. Secondly, our data indicate that the testing positions assessed (i.e., using a 0- or 30-degree knee-flexion angle) may assess different strength qualities, although these qualities are moderately associated. These tests should not be used interchangeably, and practitioners must consider the value of assessing in each position and how this may inform their practice. The lack of evidence observed to refute the null hypothesis across sEMG sites and the large variability in sEMG data indicate the need for further exploration of muscle activation during the identified isometric kneeflexor strength assessments, to provide insight into the contribution of the knee-flexors and hip extensors to each test. As suggested by Read et al. (2019), it is possible that the interpretation of instructional cues during these tests may have influenced test performance (i.e., reliability) and muscle activation, with participants potentially using different strategies to achieve maximal voluntary contraction. For example, during the ISO-Prone the wider variability in muscle activation observed in the muscles which play a more prominent role in hip extension (i.e., Glute Maximus and Adductor Magnus) suggest that some individuals may have adopted a hip extension strategy, rather than a knee-flexion strategy. Clear and consistent instructional cues should be used to mitigate variation in strategies during these tests.
Conclusion
In conclusion, peak force varies between the ISO-30 and ISOProne tests, with only moderate associations and large prediction intervals between tests. This indicates that these tests should not be used interchangeably. No differences in muscle activation of the knee-flexors and hip extensors were detected.
Acknowledgements
We would like to acknowledge the contribution of Mark Thistlethwaite for his help in the early data collection process.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Funding
The author(s) reported there is no funding associated with the work featured in this article.
ORCID
Matthew D. Wright http://orcid.org/0000-0002-4909-4162
Data availability statement
Data and code are available as part of the supplementary materials.
Author contributions
Conceptualization: Jonathan M. Taylor, Hermes Pallotta Writing (Original draft): Jonathan M. Taylor, Hermes Pallotta, Matthew D. Wright Writing (Review and editing): Jonathan M. Taylor, Matthew D. Wright, Paul Chesterton, Will Short, Hermes Pallotta Methodology/data collection: Phillip Smith, Jonathan M. Taylor, Will Short, Hermes Pallotta Data Analysis: Matthew D. Wright, Phillip Smith
Article Images (0)
×![]()
Bioz is the world’s most comprehensive AI search engine for scientific experimentation. The Bioz search engine offers researchers billions of data-driven product, technique, and protocol recommendations. Bioz taps into the latest advances in Artificial Intelligence (AI) – including Natural Language Processing (NLP), Machine Learning (ML) and Generative AI to mine and structure hundreds of millions of pages of complex and unstructured scientific papers. By harnessing the power of AI, Bioz provides researchers with an unprecedented amount of summarized scientific experimentation knowledge right at their fingertips, ultimately helping to speed up drug discovery and the rate of success in finding cures for diseases. Bioz is used by over 15 Million researchers from over 17,000 different universities and companies in 196 countries.
