Experimental Approach to the Problem

A systematic review was performed in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta Analyses) guidelines for performing systematic reviews in sports science (Moher et al., 2015).

Search Strategy

A systematic search of three databases (PubMed, Web of Sciences and SPORTDiscus) was performed to identify articles published prior to October 22, 2022. The PICO (Patient, Problem, or Population − Intervention or Exposure − Comparison, Control, or Comparator − Outcome/s) design was used to provide an explicit statement of the state of the question.

Priority was given to studies focused on the assessment of photocells. Three main groups of words were established: (1) population: “athletics”, “team sport”, “sprinter”; (2) intervention: “assessment”, “assessing”, “evaluation”, “method”, “protocols”, “test*”; (3) outcomes: “jump height”, “sprint time”, “contact time”, “running speed”. Words from different groups were combined to extract as many items as possible by adding Boolean markers. Clusters of keywords (population, intervention, and outcome) were connected with OR within each cluster and AND was used to combine the three groups: (athletics or “team sport” or sprinter) AND (assessment OR assessing OR evaluation OR method OR protocol* OR test) AND (“jump height” or “sprint time” or “contact time” or “running speed”). Additionally, the reference lists of the studies retrieved were manually searched to identify potentially eligible studies not captured by the electronic searches.

Screening Strategy and Study Selection

When the referred authors had completed the search (M.R.-G., A.M.-V. and J.P.-O.), they compared the results among themselves to ensure that the same articles were identified. Then, one of the authors (M.R.-G) downloaded the main data from the articles (title, authors, journal, date and databases) and put them onto an Excel spreadsheet (Microsoft Excel, Microsoft, Redmond, USA). Then, two authors (A.M.-V. and J.P.-O.) removed duplicates. The remaining articles were screened and checked by two authors independently (A.M.-V. and J.P.-O.) against the inclusion and exclusion criteria (Table 1). There were discussions with a third author (A.S.-L.) in the event of discrepancies regarding the selection process. Possible errata for the included articles were considered. Moreover, relevant articles not previously identified were also screened in an identical manner and further studies that complied with the inclusion and exclusion criteria were included and labelled as ‘included from external sources’.

Data Analysis

The information collected from the selected studies addressed the following aspects: (1) anthropometric characteristics of the sample (sex, age, body mass, height); (2) characteristics of the assessment protocols (location of photocells, photocell technology, measurement variables, assessment test and tools used); (3) results obtained with the tests carried out; (4) conclusions endowed with scientific rigor and objectivity, which helped in the analysis of the casuistry of the data obtained.

Quality of Studies

The methodological assessment process was performed by two authors (A.M.-V. and M.R.-G.) using an adapted version of the STROBE assessment criteria for cross-sectional research (O’Reilly et al., 2018), looking for studies that were eligible for inclusion. Each article was assessed based on 10 specific criteria (Table 2). Any disagreement was discussed and solved by consensus. Each item was evaluated using numerical characterization (1 = completed; 2 = non-completed). As suggested by O’Reilly et al. (2018), each study rating was qualitatively interpreted using the following law: the study has a risk of bias or low quality with a score lower or equal to 7 points, while those studies with higher scores are considered as of low risk of bias or high quality.

Effect Measures

The data validity was interpreted according to the concordance value of the data collected by the photocell measurement systems compared to the gold standard (Hopkins et al., 2009). For this purpose, the intraclass correlation coefficient (ICC) was assessed according to the following scale of values: <0.50 (poor), 0.5–0.75 (moderate), 0.75–0.9 (good) and >0.90 (excellent) (Koo and Li, 2016), as well as Pearson's correlation coefficient (r) analysis was conducted with the interpretation as follows: non-significant (r < 0.10), low (0.10–0.30), moderate (0.30–0.50), high (0.50–0.70), very high (0.70–0.90), almost perfect (r > 0.90) and perfect (r = 1.00) (Hopkins et al., 2009).

Identification and Selection of Studies

The database search identified a total of 164 titles (PubMed = 61; Web of Science = 48 and, SPORTDiscus = 55). These studies were then exported to reference manager software (EndNote^TM X9, Clarivate Analytics, Philadelphia, PA, USA). Duplicates (73 references) were subsequently removed either automatically or manually. The remaining 91 articles were screened for their relevance based on titles and abstracts, resulting in the removal of further 61 studies. Following the screening procedure, 30 articles were selected for in-depth reading and analysis. After reading full texts, further 14 studies were excluded due to not meeting the eligibility criteria (Figure 1).

Figure 1

PRISMA flow diagram.

Methodological Quality

The overall methodological quality of the cross-sectional studies can be found in Table 2.

Study Characteristics

Tables 3–5 summarize the characteristics of the studies, the validity of photocells for assessing sports movements and the reliability of photocells for assessing sports movements, respectively.

Factors Affecting the Validity and Reliability of Photocells for Monitoring Horizontal Displacement

It is well known that during sprinting the movement of the arms and legs is wider and faster than that of the trunk and the total center of mass, which might cause the timing gate to be prematurely or repeatedly triggered when using single beam photocells (Doma et al., 2023; Haugen and Buchheit, 2016). Altmann et al. (2018) showed that measuring horizontal displacement speed from a standing starting position with single beam photocells yielded a poor ICC (ICC = 0.278) in the first 10 m, likely because the lower running speeds and shorter time frames during the acceleration phase increased the magnitude of this error (Altmann et al., 2018; García-López et al., 2012). When starting from a crouched position, another source of error in the first meters is added because the body tends to tilt forward at least until the 13^th step (or approximately 18.5 m) (Nagahara et al., 2014) which would also trigger the photocell beam before the center of mass reaches that position. For sprints longer than 20 m or with a flying start, validity of single beam photocells has been proven excellent (Altmann et al., 2018; Bastida Castillo et al., 2017). Another factor to consider is the photocell height, for instance, they can be placed near the ground to measure the start in standing position starting sprints (Altmann et al., 2018b) or at a height varying from knee to head height to detect the runner breaking the beam. This height can be adjusted to match specific needs (i.e., different athletes’ heights, different starting positions etc.). Photocells at lower beam heights (< 0.64 m) correspond to the shortest times, due to the fact that on many occasions the light beam is cut by the thigh, being sometimes difficult to detect by dual-beam or split-beam and post-processing systems. However, it has been shown that there were no significant differences between 0.64 and 1 m single beam photocell heights for a 20-m flying start sprint, decreasing the relative error (Altmann et al., 2018b), but there were differences with a split start (standing with one foot in front of the other) (Altmann et al., 2018a). This discrepancy is attributed to the fact that, for an initial distance of 0.30 m from the first timing gate, the ankle of the back foot could pass the initial timing gate simultaneously with the hip. Therefore, placing initial timing gates below the knee could improve their accuracy (Altmann et al., 2018b).

As it has been observed, during the acceleration phase over the first 10 m, the way of starting (standing or 3-point or 4-point) is a decisive factor, and the utmost rigor should be applied when deciding on the type of start to use as they are not comparable and cannot be used interchangeably (Haugen and Buchheit, 2016). To avoid these issues, the use of dual-beam or post-processing tools has been suggested (Altmann et al., 2018b), although the sensitivity of the infrared receiver should also be considered, providing further evidence of the superior validity of dual-beam systems compared to single-beam systems when aiming to avoid false triggering produced by swinging arms or legs (Haugen and Buchheit, 2016). This factor is not generally reported, but in some measuring devices it can be adjusted so that the light beam can identify small segments such as fingers, arms or the trunk, in order to consider only the segment of interest and avoid false triggering. However, 0.60 m height dual-beam photocells produce significantly faster times than 0.80-m height photocells for 10- and 20-m standing start sprints (Cronin and Templeton, 2008). Cronin and Templeton (2008) did not report whether these heights corresponded to the upper or the lower photocell, nor the distance between the two beams. Surprisingly, we found no investigations that aimed to assess the ideal distance between both beams in a dual-beam system, although there seems to be a consensus to place them 0.2 m apart. The results show that the relative error decreases considerably from 30 m onwards, likely because the photocell beam is no longer triggered early by the forward-leaning trunk as it becomes upright after just 16 steps from a block start (Nagahara et al., 2014).

Factors Affecting the Validity and Reliability of Photocells for Monitoring Vertical Displacement

Similarly as for horizontal displacements, photocell systems have been used for vertical displacement measurement, and were generally compared to force plates with sampling rates between 500 and 1000 Hz (Attia et al., 2017; Cronin and Templeton, 2008; García-López et al., 2013; Hanley and Tucker, 2019). The tests used were a squat jump (SJ), a counter-movement jump (CMJ) and a loaded counter-movement jump (CMJ+). The inter-trial intraclass correlation coefficient (ICC) of SJ, CMJ and CMJ+ heights obtained in the studies included in this review (range: 0.93 to 0.99) was in agreement (ICC: 0.95) with that reported by Nuzzo et al. (2011) measured with a Myotest accelerometer. Furthermore, if data are homoscedastic as in the study by Attia et al. (2017), SEM analyses may be more useful to establish absolute reliability (Atkinson and Nevill, 1998). In heteroscedastic data analyses including the coefficient of variation (CV) are recommended (Atkinson and Nevill, 1998). These results obtained from different manuscripts suggest that photocells are a reliable instrument for measuring vertical jump height estimated from flight time when compared to the gold standard (force plates).

Regarding validity, the different indexed studies (Attia et al., 2017; García-López et al., 2013; Glatthorn et al., 2011; Heredia-Jimenez and Orantes-Gonzalez, 2020) show a certain degree of agreement that photocells are valid systems for measuring vertical displacement when compared to force plates (gold standard). Despite ICCs for validity very close to 1 in all manuscripts, it is important to point out that it is a system that presents a certain systematic error of underestimation (Attia et al., 2017; García-López et al., 2013; Glatthorn et al., 2011) in jump height recorded by photocells. These findings could be attributed to the number of LEDs (32 LEDs), an adequate sampling frequency (1000 Hz), as well as the height of the LED barrier with respect to the ground, because the higher the photocells are, the lower the measured flight time will be. For example, for a gravitational value of 9.81 m·s⁻² and a jump of 0.5 s, the jump distance can be calculated as 30.656 cm. Photocells with a small variation of 1 cm in height with respect to the ground would measure a time of flight of 0.49191 s, thus giving a jump height of 29.672 cm (i.e., 0.98 cm less). Despite these significant underestimates of the jump height data recorded by photocells, these systems appear to be valid and reliable, since there were significant correlations and systematic errors among the methods, and it was possible to establish prediction equations to overcome the underestimates shown by photocells versus force plates.