Neuroanatomical Correlates of Voluntary Modulation of Accommodation in the Human Visual System

Hans O. Richter(1), Joel T. Lee(2), & José V. Pardo(2,3), (1)Brain Sciences Center, VAMC, Minneapolis, MN, and Department of Physiology, University of Minnesota; (2) Cognitive Neuroimaging Unit, VAMC, Minneapolis, MN; (3)Department. of Psychiatry, Div. of Neurosience Research, University of Minnesota, Minneapolis, MN



Although accommodative eye-movements can be elicited volitionally, the brain mechanisms subserving higher cortical computations involved in VA remain elusive.

With the advent of non-invasive imaging techniques such as PET, it is now for the first time possible to explore in normal subjects the functional neuroanatomy of the Accommodative System.

Several experiments reported in the literature indicate that volitional increases in dioptric strength of the crystalline eye-lens, or VPA, can be learned rapidly and can be produced on voluntary demand without significant co-activation of vergence eye-movements. The purpose of our work was to identify human neural circuits involved in the execution of this type of VA.


PET:Positron Emission Tomography
VA:Visual Accommodation
VPA:Voluntary Positive Accommodation


1. Subjects.

7 naive subjects, 5 females and 2 males; mean age of 32 yr. with s.d. of 6yr., range 24-43; all right-handed. All subjects had normal or normal corrected vision. Subjects gave informed consent.

2. Behavioral paradigm.

The aim of the behavioral paradigm (
Figure 1 & Figure 2) was to provide: 1) a reference condition involving stable VA (condition A); 2) the same oculomotor state as in A but with a defocused retinal image (condition B); 3) continuous large amplitude (~-5.0 D), high frequency (0.5 Hz), VPA responses to a defocused retinal image (condition C).

3. Study Design and Subtractions.

Each subject was scanned during at least two blocks; each block contained three task conditions counterbalanced in order. The PET technique permits a visualization of the brains representation of the retinal image (blurred checkerboard), visualization of changes in this visual representation as brought about by VPA, and identification of VA sensori-motor process (Figure 3). To explore the different components of brain work in VPA, the following subtractions were created: 1. A-B (n.s.); 2. B-A (n.s); 3. C-A; 4. A-C; 5. C-B; and 6. B-C.

4. PET Scanning.

PET scans were acquired with an ECAT 953B camera in 2D mode for 60 sec. after the arrival of radioactivity into the brain following an intravenous bolus injection of 50 mCi H(2){15}O. Images were reconstructed using filtered back-projection to a final image resolution of 11 mm FWHM.

5. Image Analysis.

Software developed by Minoshima (Minoshima, et al., J Nucl Med, 1994) realigned scans from each subjects, estimated the intercommissural line, normalized the tissue activity and stereotactically transformed the data using non-linear warping. ANALYZE (BIR, Mayo) permitted image processing and display.


Activity Increases:
Primary and Association Visual Cortices
Superior Temporal Sulcus
(Figure 4)
Activity Decreases:
Lateral Intraperietal Area (LIP)
Cerebellum (dentate)
Frontal Eye Fields (FEF)
Supplementary Eye Fields (SEF) C-B
Left Cingulate Rostral
Left Mid-Cingulate
(Figure 5)
Null Changes:
Superior Parietal
Right Anterior Cingulate (BA 24)

Figure 1. Schematic layout of the experiment.
The checkerboard was placed 40 cm from the subjects. The "full-focus condition" (A) involved no refractive blur or dioptric changes in the AR. Subjects were requested to "look at the checkerboard naturally, the same as they would when viewing a book or a sign at the same distance".

In (B) the retinal image was defocused by a -5.0 D lens and subjects were asked to "look at the checkerboard, but to make no effort to focus on it when the lens was placed on front of the eye, I.E., to let the checkerboard be blurry".

Condition (C) involved maximum frequency and amplitude of VPA. Subjects were instructed to "look at the checkerboard and after each repeated occasion of defocus carefully focus on the checkerboard so that it is maximally sharp and clear at all times".


Figure 2. Focusing Paradigm.
When the -5.0 D lens is added to the single naked (or corrected) eye, it remains defocused for ca. 500 msec (B). After this time a "break-down" in the AR occurs which has as a result a regression of the AR to its resting-point. For a typical viewer the resting-point of accommodation averages 1.5 D. This has as a result, in the case of a typical observer, a 1.0 D incerase in net degree of defocus. At the end of the intertrial period, when the -5.0 D lens is removed from the eye, an overshoot occurs during conditon B, until the retinal image is again in focus within ca. 300 msec. During the VPA, the target is rendered clear by a voluntary efference to the ciliary muscle. The removal of the -5.0 D lens in C again results in an overshoot and error of focus of the same magnitude as the VPA response itself. In a separate behavioral session, subjects demonstrated excellent performance based upon Snellen acuity testing.

Figure 3. Sequantial task-model for the AS.
Sensory-Motor Processing

Figure 4. Cerebral regions activated during voluntary positive accommodation (C-A). Thresholded with P=0.005.

Figure 5. Cerebral regions deactivated during voluntary positve accommodation (A-C). Thresholded with P=0.005.


Increases. VPA produced activation within visual cortices (BA 17, 18, 19). Activity within visual cortices most likely reflect processing specific to VA since the C-A subtraction would not have resulted in visual activation even if the subjects' VPA responses were completely accurate. Superior temporal sulcus activation may reflect integration of visual and premotor processing. Stimulation of this region in monkeys results in accommodation responses(Jampel, J Comp Neurolgy, 1955). Insular activation, frequently observed during tasks involving autonamic processing, may mediate the integration of sympathetic and parasympathetic outflow to the ciliary muscle. As in other systems, the cerebellum may compute the difference in sensory and motor maps based upon dioptric defocus and VPA.

Decreases. Monocular VPA in response to a blurred checkerboard pattern produced significant activity decreases in a circuitry belonging to the saccadic eye-movement system. The existence of monosynaptic connections between the frontal eye-field (FEF), lateral intraparietal area (LIP) and the dentate nucleus of cerebellum has been established (Lynch et al., Exp Brain Res, 1994). FEF is known to receive input from LIP, supplementary eye-field (SEF) and the dentate nucleus of cerebellum. A decrease in activity in all of these structures was observed in our paired image subtractions during VPA. The decrease in activity in the left mid cingulate and left cingulate rostral may reflect oculomotor processes operating on the superior colliculus to inhibit saccadic eye-movement activity. Although the neural correlates to vergence remain unclear at present, it is also possible that some of the activity decreases relate to inhibition of vergence.

Null. Absence of right anterior cingulate (BA 24) activation during VPA suggests that volitional changes in VA do not recruit the anterior, executive attention system; presumably, VA processing is relatively automatic and effortless. The lack of parietal activation may reflect the absence of changes within extrapersonal space of stimuli or of attention shifts. Visual fixation neurons of the posterior parietal cortex (7a) can produce command signals to accommodation and vergence (Sakata et al., J Neurophysiol, 1980).

Copyright © 1996 Cognitive Neuroimaging Unit, VA Medical Center, Minneapolis, MN
Send correspondence to Hans O. Richter
<Cognitive Neuroimaging Unit Home Page>