Phosphenes Electrically Evoked with DTL Electrodes: A Study in Patients with Retinitis Pigmentosa, Glaucoma, and Homonymous Visual Field Loss and Normal Subjects

Abstract

purpose. To develop an improved, easy, and safe method for reliably eliciting electrical phosphenes and to compare reported perceptions and values of chronaxie and rheobase in healthy individuals and patients with distinct ophthalmic diseases.

methods. DTL electrodes were used with a four-alternative, forced-choice method to determine psychophysically the strength–duration curves by using the Weiss model of electrical tissue stimulation in 47 subjects: healthy individuals (n = 17; n = 6 under light- and dark-adapted conditions), patients with open-angle glaucoma (n = 9), retinitis pigmentosa (RP; n = 14), amblyopia (n = 3), or homonymous visual field loss (n = 4).

results. In all subjects, thresholds were safely determined without side effects. Subjects reported a homogenous, central, white, steady phosphene. The rheobase was higher in glaucoma and patients with RP than in healthy subjects (0.1129 ± 0.0314 and 0.6868 ± 0.1054 vs. 0.0383 ± 0.0057 mA; P = 0.05 and <0.00001, respectively), and it increased with age (r = 0.51; P = 0.038). The rheobase was lower in light- than in dark-adapted conditions (difference 0.015 mA; P = 0.016). In patients with RP, no correlation was found between rheobase or chronaxie and retinal thickness (in OCT) or number of years after reading loss, but there was a correlation between rheobase and visual acuity (P = 0.014). In patients with RP or glaucoma, no inhomogeneity of phosphenes was reported, but all patients with homonymous visual field loss reported lateralization of the phosphene into the area of visual loss.

conclusions. Use of DTL electrodes to elicit electrical phosphenes is safe, fast, and reliable. It bears significant advantages over corneal electrodes and provides a valuable tool to elucidate ophthalmologic disease processes and to screen candidates for retinal prostheses.

Phosphenes are defined as luminous impressions caused by excitation of the visual system by something other than the impingement of rays of light on the retina. They can be elicited by mechanical force, eye movements, chemical agents, radiation outside the visible spectrum (e.g., x-rays), magnetic fields, or electrical currents passing through the eye ¹ and have been used in basic and clinical research. ^{2 3 4 5} More recently, the use of electrically evoked phosphenes (EEPs) has increased to the examination of eyes that are to undergo implantation of a visual prosthesis ^{6 7 8 9 10} (Gekeler F et al. IOVS2006;47;ARVO E-Abstract 3193;Zrenner E et al. IOVS2006;47;ARVO E-Abstract 1538). It appears advisable to use extraocular electrical stimulation as a method of testing eyes for electrical excitability before performing invasive procedures with the risk of severe adverse advents. For the noninvasive elicitation of EEPs many different active electrodes have been used—for the most part, different contact lenses, often self-built, attached to the cornea of the subject. ^{3 4 5 6 10 11 12 13 14 15 16 17} In some cases electrodes on the periorbital skin or on the eyelid have been used for stimulation. ⁸ Corneal electrodes often have the significant disadvantage of generating unpleasant sensations and corneal problems, ¹⁰ such as pain or superficial corneal disease, and thus complicate long-term examination and interfere with detailed evaluation near the threshold. In addition, differing electrodes used for stimulation make results of different studies difficult to compare. 

We therefore propose a new system of stimulation with DTL electrodes (named after Dawson, Trick, and Litzkow ¹⁸ ), which are in common use for clinical registration of electroretinograms. DTL electrodes are easy-to-construct, single-use items and pose minimal discomfort to the patient. We have psychophysically determined thresholds for different pulse lengths to retrieve chronaxie and rheobase using the Weiss model of electrical tissue stimulation in a series of healthy individuals and patients with open-angle glaucoma, retinitis pigmentosa (RP), amblyopia, and homonymous visual field loss. 

Materials and Methods

Electrodes and Stimulation

DTL electrodes were used as active electrodes after their original description. ¹⁸ In our case, the fibers of the electrode consisted of four nylon filaments that were metallized with silver. They were received from the rug industry where they are used to remove static charge (resistance coefficient of combined filaments = 3 kΩ/m). A 4-cm piece is clamped on one side by a connector and on the other side by adhesive tape (Fig. 1A) . The latter is then fixed on the skin close to the medial canthus, and the clamp is attached to the cheek of the subject. The thread then lies on the bulbar conjunctiva just below the corneal limbus. The ground was a gold-plated cup electrode (LKC; Gaithersburg, MD) on the ipsilateral temporal skin. 

A neurostimulator (Twister; Dr. Langer GmbH, Waldkirch, Germany) was modified by the manufacturer to scale down current output. We used monopolar-positive, first-rectangular-current pulses of 0.05, 0.075, 0.1, 0.3, 0.5, 075, 1.0, 2.5, 5.0, 25.0, and 50.0 ms. 

Impedance of the electrodes was tested with an electrophysiologic testing unit (ColorBurst and Espion; Diagnosys LLC, Littleton, MA) and never exceeded 8 kΩ at 25 Hz. 

A four-alternative, forced choice-method was used for psychophysical determination of thresholds. At each stimulus duration a train of one, two, three, or four pulses of the same current strength was given and subjects had to name the number of pulses. The correct answer three times out of four was judged as “seen” or suprathreshold. The current amplitude was changed until the subject named the number of pulses correctly fewer than three times. After the threshold was determined, the stimulus duration was changed. The subject was given a prompt when the series of pulses was started. Pulses were delivered approximately every 1.5 seconds starting at 1 ms and then switching between longer and shorter pulse lengths. Subjects were seated in a completely darkened room. Full darkness was necessary to perceive the very subtle phosphenes, which was not possible in a brighter surrounding because of the lack of contrast. The room light was switched on in regular intervals for the light-adapted experiments, to avoid dark-adaptation during the experiment. In a subgroup of normal subjects, dark adaptation was reached after 30 minutes. Subjects were asked about the location, size, color, and temporal properties of the phosphene and to report any missensations such as muscle activation or pain. 

Subjects

All experiments were undertaken with the understanding and written consent of each subject, respecting the Code of Ethics of the World Medical Association (Declaration of Helsinki) and in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC). The study was approved by the local university’s ethics committee. 

Examinations were performed in 47 subjects in the following groups:

•  
  Healthy individuals (n = 17) as the control group. Six individuals in this group were also measured after dark adaptation.
•  
  Amblyopic patients (n = 3). Visual acuity was 0.05.
•  
  Patients with primary open-angle glaucoma (n = 9) with arcuate superior visual field loss (example in Fig. 1B ). The visual field center was spared, and visual acuity ranged from 0.5 to 1.0. Two patients had absolute glaucoma.
•  
  Patients with RP (n = 14). Visual acuity was either light perception (n = 9) or no light perception (n = 5). Retinal thickness was assessed by optical coherence tomography (OCT; StratusOCT 3000; Carl Zeiss Meditec, Dublin, CA).
•  
  Patients with homonymous visual field loss due to cerebral infarction (n = 4). One patient had a quadrantic loss and three had hemianopsias (example in Fig. 1 ). Visual acuity was normal in all patients.

Data Evaluation

The empiric Weiss model of electrical tissue stimulation ¹⁹ was used to perform a nonlinear fitting to determine rheobase (defined as the minimum suprathreshold current amplitude, theoretically at infinite pulse length) and chronaxie (the pulse length at twice the current of the rheobase) of each individual. Before a least-squares fit was performed, a logarithmic transformation had to be performed to adjust the residuals’ distribution. The Weiss model equation is as follows:  

\[I_{\mathrm{T}}{=}I_{0}\left(1{+}\ \frac{{\tau}_{\mathrm{e}}}{t}\right),\]

where I _T is the threshold current that leads to stimulation if offered for the time t. Rheobase I ₀ and chronaxie τ_e are considered physiologic tissue constants. 

After analysis of variance (ANOVA) the Dunnett test was used to compare the groups’ means against the mean of the control group. For comparisons between measurements in dark- and light-adapted conditions in the same individual, a paired t-test was performed. All calculations and statistical analyses were performed with commercial software (JMP IN, version 5.1; SAS Institute, Cary, NC). 

A Bland-Altman-analysis ²⁰ assessed repeatability in five subjects of the control group, to calculate the mean difference and the coefficient of repeatability (CR; calculated as 1.96 · SD of the differences). 

Results

General Remarks

All measurements were completed as described in the Methods section. No measurement had to be interrupted because of unpleasant sensations. In a few patients, a slight twitch in the lower lid leading to a minimal muscle movement occurred with the shorter stimulation pulses (0.05–0.3 ms). Occasionally, especially in diseased individuals in whom the thresholds were high, these sensations occurred at current amplitudes lower than that for the phosphene. Those measurements were not counted because of possible confusion. Pain was not reported by any individual. Ophthalmic examination failed to reveal any alterations, and diseased individuals, even at the highest thresholds, did not accept the offer of seeing an ophthalmologist for later feelings of discomfort or other ophthalmic problems. 

Healthy Individuals

All subjects in the control group reported a whitish circular phosphene in the center of the visual field extending approximately 20° in each direction. Brightness usually decreased a little with the second, third, and fourth pulses (if given). Different flash durations were not noticed. 

The rheobase was 0.0383 ± 0.0057 mA and chronaxie, 11.27 ± 2.11. The mean of all healthy individuals was fitted using the Weiss model with a high determination coefficient of r ² = 0.889. We found a linear correlation between the standard deviation of the threshold current and the logarithm of stimulus duration (SD = 0.196 + 0.066 · log(t) mA; r ² = 0.740; P < 0.01) and were thus able to plot the 95% confidence intervals for the Weiss model of the means of the group (Fig. 2) . 

Assessment of Repeatability.

The mean difference between two measurements was 1.03 ± 2.59 ms for the chronaxie and −0.008 ± 0.011 mA for the rheobase. The CRs were 5.08 ms and 0.022 mA (Figs. 3A 3B) , respectively. 

Differences in Rheobase and Chronaxie between Light- and Dark-Adapted Eyes.

We found a statistically significant higher rheobase in the dark- than in the light-adapted eyes, with a mean difference of 0.015 mA (P = 0.016). Chronaxie was higher in the light-adapted state; the mean difference was 5.2 ms (not significant). An example is shown in Figure 4 ; data are given in Table 1 . 

Age Dependence of Chronaxie and Rheobase.

The rheobase calculated in the control group showed a statistically significant positive correlation with the age of the subjects with a Pearson’s correlation coefficient of r = +0.51 (P = 0.038). Figure 5shows the rheobases of the control subjects in relation to their ages and 95% confidence intervals of the mean. There was no significant age-dependence of the chronaxie (Pearson’s correlation coefficient; r = −0.2; P = 0.4; data not shown). 

Patients with Amblyopia

The subjective description of the phosphenes did not differ from the control group. The rheobase and chronaxie (Fig. 6A)in this group were not significantly different in comparison to those in the control group (0.0505 ± 0.0063 mA, P = 0.793; 6.91 ± 0.69 ms, P = 0.958, respectively; Fig. 7 , Table 1 ). All rheobase values are within the 95% confidence interval of the control group (Fig. 5) . 

Patients with Primary Open-Angle Glaucoma.

The description of the phosphenes was not different from the control group. In particular, none of the patients reported any difference in phosphene brightness from intact areas in visual field testing and areas of scotomas. In two patients with absolute glaucoma, no phosphenes were elicitable despite stimulation with currents approximately 100 times as high as the average value in the control group. These patients were not included in the statistical work-up. 

The rheobase in this group was significantly higher than in the control group, but the chronaxie was not significantly different (0.1129 ± 0.0314 mA, P = 0.05; 16.55 ± 4.62 ms, P = 0.783, respectively; Figs. 6B 7 ; Table 1 ). Rheobase levels in three patients are within the 95% confidence interval of the control group: 1 below and 5 above (Fig. 5) . 

Patients with Retinitis Pigmentosa.

The subjective description of the phosphenes was not different from that in the control group. All patients reported an evenly distributed phosphene and none of the patients reported any correlation to remaining “islands” of visual function within the visual field (if present). In four patients with RP, no phosphenes were elicitable, despite stimulation with currents approximately 100 times as high as the average current in the control group. These patients were not included in the statistical work-up. 

The rheobase in this group was significantly higher than in the control group, whereas the chronaxie was not significantly different (0.6868 ± 0.1054 mA, P < 0.00,001; 13.81 ± 5.17 ms, P = 0.970, respectively; Figs. 6C 7 ; Table 1 ). The fitting for the means of the patients with RP does not run parallel to the fitting of the control group’s means. For the long pulses the difference is much greater and far above the upper 95% confidence interval, whereas for the shorter pulses the fitting approximates the fitting of the control group and lies well within the 95% confidence interval (Fig. 6C) . In all patients with RP, rheobase is above the 95% confidence of the control group (Fig. 5) . 

There was no correlation between rheobase and the number of years after loss of reading ability or between rheobase and retinal thickness as assessed by OCT. However, there was a statistically significant difference in rheobase between patients with and those without light perception (P = 0.014). The chronaxie did not correlate significantly with any of these values. 

Patients with Homonymous Visual Field Loss.

Most striking in this group was that all four subjects unanimously and spontaneously reported that the phosphene was brightest in the area of their visual loss. 

The rheobase and chronaxie in this group were not significantly different in comparison with those in the control group (0.02785 ± 0.0106 mA, P = 0.934; 18.59 ± 8.01 ms, P = 0.709, respectively; Fig. 6D 7 ; Table 1 ). The fitting for the means of this group runs parallel to the fitting of the control group’s means—that is, there is no difference between the short and the long pulses in comparison to the control group (Fig. 6D) . All rheobases in this group are outside the 95% confidence interval of the control group—three of them below and one slightly above (Fig. 5) . 

Discussion

Our study has shown the feasibility of using DTL electrodes to elicit EEPs in healthy individuals and patients with various visually impairing diseases. The advantages of DTL electrodes are manifold. They are easy and fast to construct, they are inexpensive, and there is no need to sterilize them after use. They can be used in cases when corneal electrodes are contraindicated ²¹ and are usually tolerated well, so that they can even be used in children of practically any age. ^{18 22} Eliciting EEPs through DTL electrodes did not cause unpleasant sensations in the eye, even when high current amplitudes were used. Only at shorter pulse lengths (0.05–0.3 ms) did subjects occasionally report a slight muscle twitch in the lower eye lid, perhaps due to stimulation of the facial nerve fibers, which is in accordance with findings that shorter stimuli tend to stimulate nerve fibers, whereas longer pulses show a preference for cell bodies. ^{23 24} DTL electrodes appeared very safe throughout our study. No morphologic alterations were noticed on ophthalmic examination. DTL fibers arguably represent the easiest conceivable electrode for electrical stimulation of the eye, minimizing contact to the eye and posing minimal constraints on the tested person. 

All healthy subjects reported homogeneous, steady, white flashes in the central visual field extending an estimated 20° in each direction from the center. This corresponds to the area of the highest rod and cone bipolar cell density in the primate retina. ²⁵ Although there is ongoing debate on the exact site of electrical stimulation within the retina and some have argued that the photoreceptors are preferentially driven, ¹⁷ evidence is emerging that photoreceptors and their synapses to bipolar cells are the preferential target because even retinas without any functioning photoreceptors can be stimulated within reasonable thresholds, ^{1 3 4 17 26 27} a finding that is supported by our study where even patients with RP with only slight light perception had near normal thresholds (e.g., patient 2 in Fig. 5 ). Simultaneous stimulation of all wavelength bipolar cells or synapses to photoreceptors would arguably yield a whitish phosphene. 

Patients with glaucoma with arcuate superior visual field loss reported the same central homogeneous phosphene as the control group, and no difference was noticed between the affected upper and the unaffected lower hemisphere. Whereas it might be difficult for subjects to describe the exact size of a phosphene, the sparing of the visual field center in our patients (Fig. 1B)could explain the intact phosphene perception in these patients and lack of correlation to the visual field loss. It was expected that patients with absolute glaucoma did not perceive EEPs because ganglion cells which are considered the primarily affected cells in glaucoma ^{28 29 30} represent the most centrally located neurons in the retina, and their loss will inhibit conduction of any signal to the brain. In patients with RP, only central phosphenes were reported, and no correlation with remaining visual islands was found. This is somewhat in contradiction to other studies, which have found phosphenes at thresholds that begin in the periphery and reach the center only at higher amplitudes, ¹⁰ although homogeneous excitation of the whole retina can be assumed through external stimulation by modeling the volume conductors of the eye. ³¹ Considerable reorganization processes and build-up of nonphysiological synapses in RP retinas ³² with misdirection of electrical excitation could be responsible for the perceived flicker phenomena in some patients with RP and seems expectable from the spontaneous observation of these phenomena by some of these patients. In contrast to the homogeneous phosphene in glaucoma and patients with RP, it is very striking that all patients with homonymous visual field loss (quadrantic or hemianopic) did not perceive their visual loss but rather lateralized the phosphene spontaneously and even with a touch of surprise in the area of their loss. It is well known that cortically blind patients—in complete absence of acknowledged awareness—unconsciously localize light stimuli into their impaired hemisphere (known as blindsight ^{33 34} ). It has been argued that extrastriate cortical areas are involved in this unconscious perception, ³³ and it may be that a comparable phenomenon in our patients led to perception of EEPs in the area of their visual loss by processing of these electrically generated signals in extrastriate areas, with bypassing of the impaired brain region to reach conscious perception. 

The values of the fitting curve of diffusely EEPs in patients with amblyopia (Fig. 6A ; Table 1 ) were practically identical with those in the control group, which is in accordance with electrophysiologic findings that show reduced cortical and retinal responses from pattern stimulation and normal responses from full-field stimulation. ³⁵  

Our finding that thresholds are significantly lower in the light-adapted state is in opposition to previous studies that showed no dependency on retinal adaptation. ^{3 16} The difference was clear, however, in each individual subject and the difference were easily appreciated by determining thresholds directly after pointing an ophthalmic flashlight at the eye and after only 30 seconds of dark adaptation. The speed of this mechanism correlates with the speed of the initial cone sensitivity loss during dark adaptation ³⁶ and indicates preferential involvement of the cone or the cone bipolar system in EEP perception. 

The rheobase and chronaxie were aptly estimated and used in the statistical work-up of the Weiss model of electrical tissue stimulation ¹⁹ in all subjects. Using a fitting model to estimate rheobase and chronaxie and evaluate electrical excitability is superior to drawing conclusions from thresholds at discrete pulse durations only: first, because the influence of possible outliers is reduced, which makes evaluation of physiological parameters more robust, and second, because the evaluation of the fitting curves produces more information than does a single threshold alone (e.g., a difference between longer and shorter pulses can be appreciated; Figs. 6B 6Cin patients with glaucoma and RP, respectively). 

Our rheobase levels are on the order of magnitude of other studies of ocular surface stimulation. Delbeke et al. ⁸ found an average value for the rheobase in 10 healthy individuals of 0.28 mA using periorbital skin electrodes, which is an order of magnitude higher than our average rheobase of 0.038 mA (calculated from the Weiss model; Table 1 ) and can be readily explained by the addition of a serial resistance by the closed eyelid. Dorfman et al., ⁵ using a Burian-Allen contact lens electrode, found thresholds of between 2 and 3 mA at 1-ms pulse duration in four healthy individuals. This threshold is approximately an order of magnitude higher than our finding of 0.3 mA (Fig. 2B) . Miyake et al., ¹⁶ using a corneal contact electrode, found thresholds in healthy individuals at a 5-ms pulse duration of 0.29 mA, which is approximately three times higher than our value of 0.096 mA at 5 ms (Fig. 2B) . It must be considered that Miyake et al. ¹⁶ and Dorfman et al. ⁵ used objective measurements of thresholds by registering cortical potentials that certainly yield higher thresholds than subjective estimation. Finally, Morimoto et al., ¹⁰ using a Burian-Allen contact lens electrode, found a threshold in healthy individuals of 0.065 mA, which is still twice as high as our rheobase of 0.038 mA. We speculate that the psychophysical determination of thresholds yields lower values than objective tests, such as recording cortical potentials or pupillary responses, such as Morimoto et al. performed. In patients with RP, values of approximately 4 mA at 4 ms are have been recorded (with a Burian-Allen corneal contact electrode), which is 1.5 times higher than our 2.63 mA at 4 ms in these patients (Fig. 6C) . Morimoto et al. ¹⁰ found a mean threshold of 0.0545 mA in patients with RP, which is somewhat lower than our rheobase of 687 mA. Both studies showed large interindividual variation (SE was 0.411 mA in Morimoto et al. and 0.105 mA in our study). Intraocular stimulation obviously requires lower thresholds, which are reported in the range of 0.1 mA at 16 ms, for epiretinal stimulation in patients with RP ³⁷ and are six times lower than our threshold of approximately 0.6 mA with extraocular stimulation (Fig. 6C) . For subretinal stimulation, no thresholds are thus far available in humans, but studies are under way (Zrenner E et al. IOVS 2006;47;ARVO E-Abstract 1538). Ultimately, it must be the goal to establish a method that allows estimation of intraocular thresholds from external stimulation, to select valid patients for retinal prosthesis implantation. 

We found no statistically significant difference in the chronaxie across groups in our study. The chronaxie and rheobase are both considered physiological tissue constants. ^{19 38} In simultaneous stimulation of a large number of different retinal cell types, axons, and polarization ^{23 24} the strength–duration curve will inevitably deviate from the original curve of single cell stimulation. The rheobase is in many studies used synonymously to the threshold current amplitude, although chronaxie is less frequently used but has been shown to be the stimulation paradigm to correlate with the least energy. ³⁸ The deviating curve through multi–cell-type stimulation where shorter pulses tend to lie closer to the normal subjects’ fitting curve (Fig. 6)could be responsible for our finding that no significant differences in chronaxie were found. 

Across all groups, the rheobase was significantly higher in our patients with RP than in the control group in analogy to all above mentioned studies. We found no correlation between the clinical parameters retinal thickness (as assessed by OCT) or years after reading loss (Table 2)and the elicitability of EEPs (as expressed by the rheobase). The only parameter that showed a correlation with the rheobase was visual acuity (light perception or no light perception; P = 0.014). However, some eyes without light perception had elicitable EEPs, and some eyes with intact light perception did not perceive them (Fig. 5 , Table 2 ). Morimoto et al. ¹⁰ found no correlation, whereas Yanai et al. ⁶ found a more distinct correlation than our study. Regarding function testing for a retinal prosthesis, none of these parameters can be used as an isolated predictor of the possibility that phosphenes can be elicited through prostheses. Because valid selection criteria for patients in implant surgery are required, it seems reasonable to use EEPs as a preoperative testing method. 

It is remarkable that the strength–duration curve in the patients with RP shows the greatest difference from the control group in rheobase level, whereas, at the shorter pulse lengths, the fitting curve lies inside the upper 95% confidence interval of the control group (Fig. 6C) . As has been mentioned, shorter stimuli preferentially stimulate axons instead of cell bodies. ^{23 24} In RP, primarily outer retinal layers degenerate, whereas inner retinal layers stay intact for a long time, ^{39 40} which would cause the curve to deviate in the way shown in Figure 6C(i.e., the estimated threshold currents tend to be nearer to those in the normal subjects at shorter pulse lengths than at longer pulse lengths). In contrast to the course in patients with RP, the curve in the patients with glaucoma (Fig. 6B)showed a tendency to be oppositely deviated (i.e., it is more distant from that of the control subjects at shorter pulse lengths). Although this difference is surely not significant, it could be explained by the preferential site of degeneration in glaucoma (which is in the ganglion cell layer ^{28 29 30} ), making a greater difference in shorter pulses likely. 

In conclusion, the use of standard DTL electrodes appeared safe and reliable throughout our study. In conjunction with a four-alternative, forced-choice method, the fitting algorithms based on the Weiss model allowed a detailed statistical work-up of the chronaxie and rheobase. The obvious advantages of DTL electrodes could help to increase the use of EEPs in research and clinics. Eliciting EEPs by the proposed method seems to be a valid predictor of the excitability of eyes that are to undergo retinal implant surgery, and valuable information can be gained in other ophthalmic diseases, as we have shown by our peculiar and unexpected finding in patients with homonymous visual field loss. 

 Supported by The Ewald and Karin Hochbaum-Stiftung.

 Submitted for publication April 24, 2006; revised July 28, 2006; accepted September 22, 2006.

 Disclosure: F. Gekeler, None; A. Messias, None; M. Ottinger, None; K.U. Bartz-Schmidt, None; E. Zrenner, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

(A) Photography of a subject with DTL electrodes attached. The DTL fibers lie on the lower eye lid just below the corneal limbus. They are connected to the positive pole of the stimulus generator with the custom-made clamp on the cheek. A gold-plated cup electrode on the temporal skin is connected to the negative pole. (B) A representative example of automated perimetry in a patient with open-angle glaucoma, demonstrating upper arcuate visual field loss with sparing of the visual field center. (C) A representative example of automated perimetry in a patient with homonymous hemianopsia with macular sparing.

Figure 2.

 

(A) Strength–duration curve in a healthy individual as determined by a four-alternative, forced-choice method with conjunctival–corneal DTL electrodes used in a double-logarithmic representation. The rheobase and chronaxie were calculated by using the empiric Weiss model of electrical tissue stimulation in a nonlinear fitting procedure. (B) Strength–duration curve in n= 17 healthy individuals in a double-logarithmic representation. Individual thresholds (points can overlap) and the average of all individuals are shown (+). The Weiss model has been used for nonlinear fitting of the means of all subjects. Shaded areas: The 95% confidence intervals for the Weiss-model of the means of the group are gray-shaded.

Figure 3.

 

Assessment of repeatability of two measurements in five subjects of the control group represented in a Bland-Altman diagram ²⁰ for chronaxie (A) and rheobase (B; CR, coefficient of repeatability).

Figure 4.

 

Strength–duration curves in a healthy individual in the control group under dark- and light-adapted conditions. The rheobase was lower in the light-adapted state (mean difference between the light- and dark-adapted state was 0.015 mA; P = 0.016; n = 6). The chronaxie was also lower in the light-adapted state (the mean difference, 5.2 ms, nonsignificant).

Table 1.

 

Averages of the Groups after Nontinear Fitting of the Average Thresholds by the Weiss Model of Electrical Tissue Stimulation

Group	n	Log of Rheobase (mA)	SE Log Rheobase (mA)	P Compared with Normal	Mean Chronaxle (ms)	SE of Chronaxle (ms)	P vs. Normal	Comments
Normal	17	−1.53 (0.0298)	0.09	—	11.27	2.11	—	—
Amblyopia	3	−1.30 (0.0497)	0.05	0.793	6.91	0.69	0.958	—
Glaucoma	9	−1.09 (0.0810)	0.16	0.05	16.55	4.62	0.783	In 3 patients with absolute glaucoma, no phosphenes were elicitable.
RP	14	−0.27 (0.5389)	0.13	<0.00001	13.81	5.17	0.970	In five patients no phosphenes were elicitable.
Homonymous anopsias	4	−1.66 (0.0218)	0.18	0.934	18.59	8.01	0.709	All patients perceived brightest phosphenes in the area of the visual loss.
Normal light-adapted	6	−1.95 (0.011)	—	0.016	17.93	—	0.060	Probabilities are given for a paired t-test light- vs. dark-adapted.
Normal dark-adapted	6	−1.60 (0.025)	—		12.69	—

 In brackets, actual values of rheobase are given as calculated from the logarithmized values. Probabilities were calculated using ANOVA and the Dunnett test; bold numbers are significant.

Figure 5.

 

The rheobase in all 47 patients and control subjects in relation to age (+, control; *, RP; ○, amblyopia; □, visual field loss from open-angle glaucoma; •, homonymous visual field loss). Shaded area: upper and lower 95% confidence intervals in the control group. The numbers next to patients with RP relate to those in Table 2 .

Figure 6.

 

Strength–duration curves in patients with amblyopia (A; n = 3), open angle glaucoma with arcuate visual field loss (B; n = 7), RP (C; n = 10), and homonymous visual field loss due to cerebral infarction (D; n = 4). Squares mark the individual values of the group; asterisks mark the average of the group. The control group is shown in each diagram, including 95% upper and lower confidence intervals (shaded area).

Figure 7.

 

Distribution of rheobase and chronaxie of all patients. Black points (jittered) are the individual levels in the different groups. Diamonds: the group distribution; the mean is represented by the horizontal line in the center of the diamond and its vertical extremities represent the lower and upper 95% confidence intervals. The smaller horizontal bar is an overlap mark. Overlap marks in one diamond that are closer to the mean of another diamond than that diamond’s overlap marks indicate that those two groups are not significantly different. The width of the diamonds is proportional to the group size. Data are given in Table 1 .

Table 2.

 

Rheobase and Chronaxie and Clinical Data of All Patients with Retinitis Pigmentosa

ID	Age (y)	Rheobase (mA)	Chronaxie (ms)	Visual Acuity	Years after Last Reading Ability	Retinal Thickness in OCT (μm)
1	70	0.84372	3.74468	LP	24	150
2	26	0.06281	30.39735	LP	8	150
3	71	1.00000	25.00000	LP	7	120
4	66	1.00000	49.99952	No LP	15	275
5	56	0.90000	11.11111	No LP	12	200
6	50	0.77177	0.38267	LP	20	180
7	48	0.91653	3.14203	LP	38	260
8	68	0.62720	3.13389	LP	20	180
9	44	0.16946	10.41979	LP	2	230
10	43	0.57617	0.79483	LP	20	165
11	56	>4	—	No LP	23	NA
12	71	>4	—	LP	6	200
13	59	>4	—	No LP	30	250
14	37	>4	—	No LP	7	130

The authors thank colleagues in the hospital who recruit the patients who participated in the study.