In this study, the researchers will investigate whether the E-effect is present in stroke subjects both on the subjective visual and postural vertical test. In addition, researchers will investigate the impact of the loss of different sensory information sources on the sensory reweighting strategies and perception of verticality by measuring the effect of somatosensory loss on the E-effect.

Postural control emerges from the interaction between the task, the environment and the individual. Within the individual, an efficient interaction between motor, sensory and neural systems is needed in order to maintain postural control.(1) One of the neural processes is the integration of afferent information such as visual, vestibular and somatosensory input to construct a body-centred frame of reference in the gravitational environment.(2) This frame of reference must be vertically aligned with the gravitational vector to ensure axial extension of the body keeping the centre of gravity within the base of support.

Each of the sensory modalities has a relative load contribution in the estimation of verticality depending on the task and the environment.(3) In healthy subjects, when certain modalities are absent or information seems biased e.g. walking in complete darkness, more weight is given to other sensory input sources.(1) This reweighting of sensory information is therefore key in postural responses in humans.

However, sensory reweighting is not always adequate when certain modalities are biased. In verticality perception, when the roll tilt of the head is less than 60°-70°, a contralateral deviation of the subjective vertical has been reported.(4) This is called the E-effect and has been observed in both the Subjective Visual Vertical test (SVV) and the Subjective Postural Vertical test (SPV).(5, 6) Furthermore, studies have shown that also somatosensory loss has a negative impact on perceiving verticality in patients after stroke.(7) In our opinion, when less afferent input sources are present, adequate estimation of the earth vertical will be more difficult. Since, studies show that verticality perception is highly related to postural control(8, 9), this is of major importance. In neurological conditions such as stroke sensory input sources are often affected, leading to fewer options in sensory reweighting strategies.(10) In this study, the researchers will explore verticality perception and sensory reweighting strategies in stroke subjects. At first, the researchers will investigate whether the E-effect also occurs in our sample of stroke subjects. Secondly, they will investigate the effect of somatosensory loss on the extent of the E-effect. It can be hypothesized that when patients have no sensory loss, more secondary afferent input is available to improve estimation of the vertical and therefore less misleading by the head-on-body tilt.

Patients and methods Study design A cohort study was designed to investigate whether the E-effect occurs in people after stroke. In addition, the effect of somatosensory loss on the extent of the E-effect will be investigated to provide further insights in the sensory reweighting strategies. Ethical approval was given by the ethical committee with registration number B300201630358 in accordance with the Declaration of Helsinki 1975, revised Hong Kong 1989 Patients Patients were recruited from the stroke population of the rehabilitation hospital Revarte, Antwerp, Belgium. All patients with a history of first stroke attending a rehabilitation program were eligible for inclusion. Patients who had an age above 80; other neurological and orthopaedic impairments as well as brainstem, cerebellar or multiple lesions were excluded. Only strokes with an ischemic or hemorrhagic etiology were included. Patients were also excluded when the subjects had pre-existing co-morbid conditions that may affect vision and somatosensory function. In addition, patients with visuospatial neglect and pusher behaviour were also excluded as this can affect verticality perception. This was examined by a neuropsychologist and the use of the Scale of Contraversive Pushing (SCP)(11). In addition, patients had to perform the assessment within three months post-stroke. Prior to inclusion, the participants were asked whether they understand the instructions of tests and to sign a written informed consent.

Outcome measures Rivermead assessment for somatosensory Performance The Rivermead Assessment for Somatosensory Performance (RASP) measures different somatosensory modalities on the face, hands and feet and has been noted to be a reliable and standardized assessment. Six tests are administered on each of the ten (five left and five right) test regions on the face; hands and feet, two tests are administered on only the face and palm of the hands. During testing eyes of the participants are closed. These eight tests can be divided into six primary and two secondary tests of sensation. Six trials are executed on each of the ten test regions, for two of the tests sham trials were given. Sham trials increase the patient's internal reliability. Patients were excluded from the statistical analysis if they had more than five false positive replies, suggested by Winward et al.(12) Subjective Visual Vertical (SVV) The Difra Vertitest type D107201 (Difra, Welkenraedt, Belgium) was used for SVV assessment. The device has an accuracy of 0.1°. A laser bar is projected at a distance of 2.5m on an opposing wall and on an altitude of 1.5m. The patients are seated in front of the device on a chair without any arm- or backrests. Patients with no adequate sitting balance were assessed while seated in their wheelchair. The room was darkened and five minutes of waiting period was given allowing the subject to adjust the darkness. Both researcher and participant obtained a remote to allow rotating the laser bar either clockwise (right) or counter clockwise (left). The researcher's remote showed a display with the amount of deviation in relation to the earth's gravitational vector. The researcher made the laser bar invisible and rotated it in a specific angle in relation to the earth vertical. Subsequently, the line was shown after which the patient had to place the line in upright position again with his nonhemiplegic hand on the remote control. The amount of deviation of each starting roll position was different for each trial. A specified order was followed: first the line was placed in 20° counter clockwise, 10° clockwise, 5° counter clockwise and 0° according to the earth vertical, followed by 5° clockwise, 10° counter clockwise and finally 20° clockwise. This series was executed three times. During the first series the patient was asked to hold the head in normal upright position, followed by a series with the head tilted to the left (while the head was bent the subjects needed to keep their trunk upright) and finally a series with the head tilted to the right side. The clockwise rotation is shown positively and the counter clockwise negatively. The patients did not receive any feedback about their performance during assessment.

Subjective Postural Vertical (SPV) The rotation chair works on hydraulic pumps and has a height of 1m. On the back of the chair, a Mitutoyo digital protractor pro 3600 (Belgium) was mounted. This allowed measurement of the deviation in relation to the earth vertical with an accuracy of 0.01°. Both the researcher and patient were given a remote to rotate the chair clockwise (right) and counter clockwise (left). Movements were restricted in the frontal plane. Before the assessment started the patient was blindfolded, depriving the subjects of visual information when readjusting the chair to earth vertical. The researcher rotated the chair as in the procedure of SVV (starting roll position of the chair). The head-on-body position is similar as in the SVV procedure. The subject had to place the chair in upright position again by placing the seating surface of the chair horizontal. The patient used his non-hemiplegic hand on the remote control. The clockwise rotation is shown positively and the counter clockwise rotation negatively.

Inclusion criteria:

• Adult first-stroke patients
• Maximum 3 months post-stroke at inclusion
• Patients attending a rehabilitation program in the Rehabilitation Hospital Revarte, Belgium

Exclusion criteria:

• An age above 65
• Neurological, except stroke, and orthopaedic impairments influencing outcome measures
• Brainstem, cerebellar or multiple lesions
• An ischemic or hemorrhagic etiology
• Pre-existing co-morbid conditions that may affect vision and somatosensory function

stroke patients within 3 months poststroke hospitalized in a Belgian Rehabilitation Hospital