Computerized Dynamic Posturography (CDP) Sensory Test Scoring -- The PREF score.

Timothy C. Hain, MD • Page last modified: July 15, 2022Return to testing indexReturn to Sensory analysis page

In these pages we are taking a deep dive into the method of scoring the Neurocom CDP sensory tests.

Processed CDP scores
Score Formula Compared Apparent Intent
Pref (3+6)/(2+5) 2 tests with stabilized vision)/(Similar tests with no vision) Distorted Vision is worse than no vision ? Distortion Vision


Typical Sensory Analysis

The PREF score is hard to understand but appears to be constructed to determine if a person is more unsteady with distorted vision than no vision at all. This it might seem to be a potential measure of "visual dependence". On the above graphic, a rather typical sensory analysis is shown for an older dizzy patient. The "Pref" score is to the far right.

Low PREF score

For illustration, the analysis above is from an individual who had vestibular neuritis one year ago, and is now complaining of being made unsteady by complex visual stimulli. There are multiple low scores, of which Pref is one of them. The score is not especially low (see norms below), and there are quite a few other low scores as well. The "VEST" score can be explained by the previous vestibular neuritis. The "Vis" score however, is even lower. So one would expect both a low VEST score and a low PREF score.

The "PREF" score computes the ratio of sway scores elicited by distorted vision (3, 6) compared to matched tests with no vision (2, 5). This appears then to be designed to quantify the degree to which distorted vision impairs stability to a greater extent than no vision. This is a messy measure depending on messy inputs -- the distortion of vision and ankle is not a stable variable from test-test, but rather depends on how much the person succeeds in staying relatively upright, as sway on the platform triggers visual surround movements. The visual distortion also depends on how well people follow instructions and look at the surround rather than at the floor. This problem could be called "GIGO" --garbage in, garbage out.

One would expect that people who score low on the PREF, would sway more with distorted vision than no vision. This could demonstrate that they have some degree of "visual dependence". This might mean that thay won't or can't switch to a more reliable input when vision is unreliable.

A relatively high PREF score (see graph of norms below), would suggest that people benefit more from distorted vision than most people. Or in other words, that visual, distorted or not, helps them balance. This line of thought makes one wonder if there is any validity to the PREF score.

For persons with normal vestibular and proprioceptive input, a low PREF score would indicate that they just won't switch to vestibular or proprioceptive (i.e. a central issue). For people with abnormal vestibular and proprioceptive input, it might indicate that they can't switch. Thus, one would expect that the "PREF" score would have meaning mainly when the proprioception and vestibular input is normal, as otherwise one gets into this conflict between won't or can't.

One might drill down more on this by considering the subcases such as 3/2 (where proprioception is available), and 6/5 (where proprioception is not available). Practically however, we do not know of anyone doing this.

Below are the group data for PREF scores. --PREF scores with respect to age in a broad group of dizzy patients.

CDP Pref scores in dizzy patients
Pref vs age
The graph above shows mean Pref scores as a boxplot from almost 2000 dizzy patients tested at Chicago Dizziness and Hearing. Graph made with "R"


Discussion of the PREF score in CDP

Age through the 7th decade is associated with a modest reduction in PREF scores. The moderate sized error bars show that PREF score is not especially variable in a "Dizzy" population. Of course, these measurements are not from "normal" subjects, but rather a mixture of many subjects with individual conditions. So some of the variability is likely due to subject composition.

Different than the other sensory analysis scores, the PREF score has many high outliers. This likely reflects the greater scatter involved in using 4 scores to derive a number rather than 2 as is the case for the other subscores, but also raises the possibility that perhaps the score just has no physiological meaning.

There have been very few academic articles examining the PREF score .

Perez et al, 2004 wrote: "The results of the caloric test in patients were independent of vision preference although canal paresis was more frequently abnormal in patients without visual preference. No differences were found in the results of rotatory stimulation by means of impulse and sinusoidal tests, both at high velocities of stimuli, in between patients with and without vision preference. Similarly, the responses in the DHI, a common questionnaire for vestibular disability and handicap and, specifically to questions addressing the problem of visual and vestibular disability, were not able to differentiate either group of patients. Nevertheless, we have found that patients with vision preference tend to have poorer balance. CONCLUSION: We consider that in the patients studied here, vision preference must be considered as a normal finding as this represents a normal strategy in a subject that relies more heavily on visual cues for his or her postural control." This would suggest that the PREF score does not correlate with most measures of vestibular function. Perez et al (2004) suggested that it is a "normal strategy". They did not suggest it was useful for visual dependence.

Parietti-Winkler et al (2008) wrote about the PREF score in patients with vestibular schwannoma. They divided up patients with low reliance on vision (G1) and high reliance on vision (G2). They reported " In G1 patients, little static posturographic and SOT performance deterioration after uVD was observed, despite vestibular test and SVV modifications. In G2 patients, uVD-related modifications followed a time-course characterised by a degradation in posturographic and SOT, vestibular and SVV performances immediately after uVD and a progressive restoration and even improvement 1 month and particularly 3 months after surgery. ". This work suggests that a high PREF score predicts a high impact of vestibular loss on balance. This suggests that it might have some role in establishing prognosis prior to intentional vestibular lesions (such as vestibular schwannoma surgery, low-dose gentamicin for Meniere's, or gamma-knife).

As the PREF score involves manipulated vision, one would expect that the newer Bertec version of CDP would greatly affect the "PREF" score.

References (for all posturography pages)