Evaluation of subjective vertical perception among stroke patients: a systematic review

Abstract Background: Verticality misperception is relatively common among patients after stroke, and it may be evaluated in terms of (a) subjective visual vertical (SVV), (b) subjective haptic vertical (SHV) and (c) subjective postural vertical (SPV). To better understand these assessment methods, we conducted a systematic review of the methodological characteristics of different protocols for evaluating SVV, SHV and SPV among individuals after stroke. Objective: To standardize the methodological characteristics of protocols for evaluating verticality perception after stroke. Methods: We searched the following databases: PUBMED, regional BVS portal (MEDLINE, LILACS, IBECS, CUBMED, Psychology Index and LIS), CINAHL, SCOPUS, Web of Science, Science Direct, Cochrane Library and PEDro. Two review authors independently used the QUADAS method (Quality Assessment of Diagnostic Accuracy Studies) and extracted data. Results: We included 21 studies in the review: most (80.9%) used SVV, eight (38.1%) used SPV and four (19.0%) used SHV. We observed high variability in assessments of verticality perception, due to patient positions, devices used, numbers of repetitions and angle of inclination for starting the tests. Conclusion: This systematic review was one of the first to explore all the methods of assessing verticality perception after stroke, and it provides crucial information on how to perform the tests, in order to guide future researchers/clinicians.


inTROdUCTiOn
Stroke can lead to multiple systemic impairments, including sensory, perceptual and cognitive disabilities. All of these can affect balance and interfere with perceptions of verticality 1 .
Verticality can be perceived in different manners: 1) visual perception of the vertical, evaluated by means of a subjective visual vertical (SVV) test that relies on visuo-vestibular information; 2) postural perception of the vertical, measured through a subjective postural vertical (SPV) test derived from graviceptive-somesthetic information; 3) tactile or haptic vertical sense, assessed through a subjective haptic vertical (SHV) test 2 .
The SVV is responsible for an individual's ability to determine whether objects are aligned vertically without a visual reference point for verticality. It depends on the interaction of sensory information in the visual and vestibular systems. This type of perception is associated with peripheral vestibular information and with integration in the parietoinsular vestibular cortex and the superior temporal gyrus 3,4 . The SHV results from somatosensory stimulation during manual exploration of an object in space 5 . The SPV reflects the capacity of the body to adjust to the gravitational vertical 6 . In the last two types of perception, the temporo-parietal junction (TPJ) is a target region for multisensory integration and processing 7 . However, it needs to be emphasized that impairment of one type of perception of verticality does not necessarily imply that the other types have been compromised.
Regarding the ontogenetic process, verticality is a significant acquisition during motor development, and it is pivotal to learning to maintain an upright standing posture, in association with standard trunk control, which is the basis for motor abilities such as reaching and gait 8 . Stroke is the main cause of impaired verticality in adults, and this impairment impacts on balance recovery, postural deficits, functionality and limitations to activities of daily living 9,10 . Studying human verticality after stroke with the aim of establishing the main assessment methods can provide the best strategy for future treatment in this population.
With this assumption, some questions still need to be clarified: 1) Are there any validated methods for evaluating verticality perception after stroke? 2) How are these tests performed? To better understand these assessment methods, we conducted a systematic review of the methodological characteristics of different protocols for evaluating SVV, SHV, and SPV among individuals who had suffered stroke.

Search process and articles selected
This systematic review followed the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 11 . However it was not possible to perform a systematic review of diagnostic test accuracy (DTA) because there is no gold-standard verticality test. Two investigators independently searched for all articles written up to November 2020, without restriction of dates or languages, using the following databases: PUBMED, regional BVS portal (MEDLINE, LILACS, IBECS, CUBMED, Psychology Index and LIS), CINAHL, SCOPUS, Web of Science, Science Direct, Cochrane Library and PEDro. Both researchers also searched through the reference lists of all studies selected.
We included original studies reporting empirical data from evaluations on perceptions of verticality among post-stroke patients. We included studies that assessed verticality perceptions among patients diagnosed with any type of stroke (ischemic or hemorrhagic), from the acute phase (in the first 24 to 72 hours) to its chronic phase. We excluded studies involving animal models, duplicate studies, systematic reviews, off-topic studies, editorials and commentaries not reporting empirical results.
From each study, we collected data on the participants (sample size, age and gender), type and time of stroke and methodological evaluation. We assessed the methodological quality of the studies using QUADAS (Quality Assessment of Diagnostic Accuracy Studies) 12 . This is a validated evidence-based tool for quality assessment that is used in systematic reviews to decrease the risk of study selection bias and increase the accuracy of the conclusions drawn from the review. It contains 14 questions, and we used the adapted protocol described by Conceição et al. (2018) 13 , which contains six of these questions: #1-Was the spectrum of participants representative of participants who will receive the test in practice?; #2-Were participant selection criteria clearly described?; #5-Did the whole participant sample or a random selection of the sample receive verification using a reference standard for diagnosis or, at least, was it confirmed verbally that the sample did not have any disease?; #9-Was the execution of the reference standard described in sufficient detail to permit its replication?; #12-Were the same clinical data available when test results were interpreted as would be available when the test is used in practice?; and ( f) #13-Were uninterpretable/intermediate test results reported? Each question was scored as 0 = study did not present this item (N) or the item is not clear in the study (NC), or as 1 = the study presented this item.
Here, we report on data from these studies descriptively, regarding the methodologies used for assessing subjective perceptions of verticality and the results obtained. We did not perform any meta-analysis because of the heterogeneity of the assessment methodologies used in these studies.

ReSULTS
We identified 1759 studies, and after reading their titles and abstracts, 26 studies were selected for further examination.  Among these, we excluded seven studies because they did not meet the inclusion criteria. Thus, we included 21 studies involving 806 participants that evaluated post-stroke patients' subjective vertical perceptions ( Figure 1). Data extracted from these 21 studies are displayed in Table 1. The participant sample sizes ranged from 5 to 86 and their ages from 43 to 76.5 years. The participants included had suffered ischemic and hemorrhagic strokes. The time that had elapsed from the stroke to the evaluation on verticality perceptions ranged from 19.4 days to 4.7 years. Twelve studies included participants who were in the subacute stage of stroke; three studies included participants in the acute phase; two included participants in both the acute and the chronic phase; and one study included participants in both the subacute and the acute phases. Three studies did not report stroke phase. Table 2 shows the QUADAS results. Among the 21 studies included in this review, most (18; 80.9%) using the evaluations of SVV perceptions 1,10,14-29 , eight (38.1%) used SPV 9,10,19,22,27,29,30,31 and four (19.0%) used SHV 9,20,22,25 .
In all of these studies, the evaluation was done in a dark room in a seated position, except for Reinhart (2016), who evaluated the subjects in an orthostatic position 23 . Nine studies involved head, chin and trunk stabilization [14][15][16][17][20][21][22][23][24]   To examine both egocentric and allocentric space representation and weightbearing asymmetry among chronic stroke patients. N/0: study did not present this item; P/1: the study presents this item; NC/0: the item is not clear in the study; QUADAS questions: (1) Was the spectrum of participants representative of the participants who will receive the test in practice? (2) Were selection criteria clearly described? (5) Did the whole sample or a random selection of the sample receive verification using a reference standard for diagnosis or, at least, was it confirmed verbally that the sample did not have any disease? (9) Was the execution of the reference standard described in sufficient detail to permit its replication? (12)  To clarify the differences in SPV with regard to the measurement starting position between patients with and without PB.  were not placed on the ground 19 . Barra (2008) 14 and Perennou (2008) 22 kept their subjects' lower limbs in a fixed position. The other studies did not report the position of the lower limbs.

SVV
Regarding the devices used for performing the tests, many studies used specific software with a light line in a dark setting or with background changes 17,23 .
In performing the test, Kerkhoff (1999) did not describe the initial angulation 20  The number of repetitions in these assessments ranged from 6-10 times, and the authors explained that half of the repetitions were in the counterclockwise direction and the other half in the clockwise direction 1,15,16,[18][19][20]23,26,28 . We can also highlight that the movement speed of the light in the SVV test ranged from 1 to 5°/s 28,31 .

Subjective haptic vertical
Four studies evaluated the SHV in this review 9,20,22,25 . Of these, two studies, Kerkhoff (1999) 20 and Utz (2011) 25 , compared individuals with and without unilateral spatial neglect (USN) with a healthy group. Perennou (2008) 22 evaluated individuals who had suffered a stroke in both hemispheres with a healthy group, and Baggio (2016) 9 also analyzed individuals who had suffered a stroke in both hemispheres. Kerkhoff (1999) 20 evaluated the SHV with different levels of head tilt, and Utz (2011) evaluated the SHV in two planes, frontal and sagittal.
In all of these studies, during the evaluations, the individuals were blindfolded in a seated position. In three studies 9,22,25 , the authors performed stabilization of the head, trunk and lower limbs; and in one, they stabilized the chin and head 20 . The length of the bar ranged from 15 to 40 cm, and the distance from the individual to the device ranged from 0.4 to 0.5 m.
The initial angle for performing the SHV test ranged from 15-45°2 2 or was 20°2 5 or 15°9. Therefore, overall, the angulation ranged from 15 to 45°. In most of the studies, ten repetitions were performed: half of them clockwise and the other half counterclockwise.

Subjective postural vertical
Eight studies evaluated the SPV in this review 9,10,19,22,27,29,30,31 . Four studies evaluated patients with Pusher syndrome 19,29,30,31 . In the evaluations, individuals were positioned seated on a tilting chair with head, trunk and lower limbs stabilized; and were blindfolded. The exception to this was Bergmann et al. (2016), who determined the SPV while their subjects were standing 30 . Other authors described positions of the lower limbs in which they were not placed on the ground 9,10,27,29,31 , while Perennou (2008) 22 described lower limbs supported on the ground.

diSCUSSiOn
This systematic review was one of the first to discuss and explore the methods of assessing verticality perceptions among post-stroke individuals, and it provides crucial information on how to perform the tests in upcoming studies, in order to standardize the evaluation.

Subjective visual vertical
In all studies, the evaluation of SVV perceptions took place in a dark room with blindfolded individuals positioned on a chair. In most of these studies, the head, chin and trunk were stabilized in the vertical position. Recently, Molina et al. (2019) showed that the type of protocol used for determining the SVV did not have any effect 32 . This suggests that there is no difference in estimating the SVV between use of an upright standing position and use of a sitting position.
The methodological differences between the studies were mainly related to the length of the luminous stimulus, the distance between it and the individual evaluated, and the initial angulation. The number of repetitions ranged from 6-10 times, and the individuals indicated the vertical position through verbal command in the majority of the studies. Utz (2011) 10 and Saj (2005) 24 used visual and proprioceptive information. Most studies evaluated SVV among individuals with USN after stroke 20,24,17,28,29 , with balance impairment 1,15 and with balance recovery 16 , and individuals with Pusher syndrome 18,19,21,22,29 .

Subjective haptic vertical
To evaluate the SHV, a bar was used in some studies, and the individuals were seated on a chair and blindfolded. There were methodological differences among the studies, such as the length of the bar, the distance from the individual to the device and the initial angulation of the test. These studies indicated that assessment of SHV perceptions should be done among individuals with USN 9,20,22 .

Subjective postural vertical
Certain factors may interfere with evaluation of SPV perceptions, such as the possibility of head and trunk movements and the somatosensory input. There were methodological differences among the SPV tests used, such as the subjects' positioning, whether their lower limbs were on the ground and the chair speed 9,20,22 . SPV tests were also used among individuals with Pusher syndrome after stroke 9,20,22,31 , and with USN after stroke 29 .

Limitations and clinical applications
Although all of these studies evaluated individuals' poststroke condition, the samples were very heterogeneous regarding the number of participants and their ages, brain lesions and length of time since the stroke. However, after analysis of all the studies included, we were able to discern some consistent tendencies regarding the methodology for carrying out these evaluations, including, for example, placing the examinee in a sitting position. The evaluations on SVV perceptions took place with a light stimulus in a dark room, and individuals indicated the vertical position through a verbal response. In evaluations on SHV, blindfolded individuals had to adjust a bar to an upright position, and in evaluations on SPV, blindfolded individuals adjusted themselves to the vertical position in a chair that had previously been rotated passively. Most studies conducted evaluations on verticality perceptions among patients with USN, Pusher syndrome and balance recovery after stroke.
In this review, we highlighted that most of the studies had a negative aspect in questions 1 and 13 of the QUADAS tool. The sample was not always representative, and uninterpretable/ intermediate test results were not well reported. According to the adapted QUADAS protocol from Conceição et al. (2018) 13 , considering these six questions, the studies with better methodological quality were the following:   16  It was clear that there was high variability in the methodological aspects of evaluations on SVV, SHV and SPV. The initial angulation of the test, the device used and the patient position require standardization. We found three systematic reviews on the methodological approaches used in verticality tests after stroke 13,32,33 . However, only our review has investigated the use of all approaches to verticality perceptions after stroke. Therefore, we have put forward an evaluation protocol for SVV, SHV and SPV to help future researchers/clinicians, which is presented in Table 3. Nonetheless, in the light of the complex development of new strategies for motor control among acute and chronic stroke patients, evaluations on verticality perceptions should be adapted according to the stroke phase. We acknowledge that there are different limitations to evaluations on verticality perceptions in the acute phase of stroke. First of all, patients in the acute phase may be clinically unstable, may have poor trunk control and may have difficulty in understanding the test. Because of this clinical situation, the patient position needs to be adapted ( for example, through use of a semi-recumbent position). In addition, it may be necessary to test patients in bed. Furthermore, verticality devices need to be easy to handle and portable.
In conclusion, we conclude that there is high variability in assessments on verticality perceptions, due to patient positions, devices used, numbers of repetitions and initial angles for starting the tests. Therefore, we have proposed a protocol for assessing each type of verticality, based on the similarities found between the studies included in this review. Furthermore, assessment of verticality perceptions is a valuable tool for cases of USN, Pusher syndrome and balance disorders after stroke.