- Email: [email protected]
Spreading and synchronization of intrinsic signals in visual cortex of macaque monkey evoked by a localized visual stimulus
Brain Research 985 (2003) 13–20 www.elsevier.com / locate / brainres
Research report
Spreading and synchronization of intrinsic signals in visual cortex of macaque monkey evoked by a localized visual stimulus Guang Bin Liu a , Ying Zhang b , John Douglas Pettigrew a , Wei Feng Xu b , Chao-Yi Li b , * a
Vision, Touch and Hearing Research Centre, Faculty of Biological and Chemical Sciences, The University of Queensland, Brisbane, Australia b Institute of Neuroscience, Shanghai Institute of Biological Sciences, Chinese Academy of Sciences, Shanghai 200031, China Accepted 22 May 2003
Abstract Spatio-temporal maps of the occipital cortex of macaque monkeys were analyzed using optical imaging of intrinsic signals. The images obtained during localized visual stimulation (IS) were compared with the images obtained on presentation of a blank screen (IB). We first investigated spontaneous variations of the intrinsic signals by analyzing the 100 IBs for each of the three cortical areas. Slow periodical activation was observed in alternation over the cortical areas. Cross-correlation analysis indicated that synchronization of spontaneous activation only took place within each cortical area, but not between them. When a small, drifting grating (28328) was presented on the fovea, a dark spot appeared in the optical image at the cortical representation of this retinal location. It spread bilaterally along the border between V1 and V2, continuing as a number of parallel dark bands covering a large area of the lateral surface of V1. Cross-correlation analysis showed that during visual stimulation the intrinsic signals over all of the three cortical areas were synchronized, with in-phase activation of V1 and V2 and anti-phase activation of V4 and V1 / V2. The significance of these extensive synergistic and antagonistic interactions between different cortical areas is discussed. 2003 Elsevier B.V. All rights reserved. Theme: Sensory systems Topic: Visual cortex: striate Keywords: Optical imaging; Visual cortex; Macaque monkey
1. Introduction The delineation of visual cortical areas by electrophysiological investigations is often problematic due to the long time span needed to map many cortical sites in order to define area boundaries. In addition, the anatomical corroboration of the thus-proposed boundaries comes too late to influence the course of a physiological experiment. Intrinsic signal optical recording, which is based on changes in local blood supply and activity-related metabolism [2,10,15], allows the functional delineation of the boundaries between relatively large areas of visual cortex within the time frame of one recording experiment. It also *Corresponding author. Tel.: 186-21-5492-1777; fax: 186-21-54921778. E-mail address: [email protected] (C.-Y. Li). 0006-8993 / 03 / $ – see front matter 2003 Elsevier B.V. All rights reserved. doi:10.1016 / S0006-8993(03)03049-X
provides a high spatial resolution and reveals the dynamic properties of cortical activation. In the current experiment, we report a novel observation using this technique. In response to a small visual stimulus presented on the fovea of the monkey’s retina, a dark spot appeared in the optical image at the area of the visual cortex corresponding to the retinal location of the stimulus. This area of activation then spread to include the greater part of the occipital cortex. Using cross-correlation analysis, we were able to study the dynamics of this extensive activation across different visual cortical areas. In addition, we observed functional boundaries between recognized visual cortical areas that were based on differences in the timing of spontaneous oscillations in the intrinsic signal recorded optically from different visual cortical areas at the same time. While further work will be needed to define the source of these spontaneous oscillations, the fact that they
14
G.B. Liu et al. / Brain Research 985 (2003) 13–20
are coherent within one area but differ across areas offers a new basis for defining various cortical areas. The present finding also suggests alternative mechanisms to explain how the boundaries are set up in the first place.
2. Methods Data were obtained from the occipital cortex of two anesthetized and paralyzed rhesus monkeys (Macaca mulatta). The methods used for the preparation of the animals were basically those which have been used for previous neurophysiological investigations in our laboratory [12]. Briefly, the monkey was initially anesthetized with ketamine hydrochloride (30 mg per kg body weight). Subsequently, the monkey received an infusion of gallamine triethiodide (10 mg / kg / h), D-tubocurarine chloride (0.25 mg / kg / h), urethane (20 mg / kg / h) and glucose (200 mg / kg / h) in Ringer’s solution (3 ml / h), for muscle relaxation, anesthesia and appropriate physiological condition. End-tidal CO 2 , body temperature and ECG were monitored throughout the experiment and these parameters were maintained within the normal range. Access to the visual cortex was provided by a craniotomy, 22 mm in diameter, so that areas V1, V2, and V4 were exposed (Fig. 1) and the dura was opened. In order to minimize cortical movement due to respiratory and cardiac activity, a stainless steel chamber system was glued to the skull, using dental acrylic. The chamber was filled with silicone oil and closed with a glass window. The exposed cortex was illuminated with light guides, using a 540-nm wavelength interference filter (bandwidth of 30 nm). Signals
obtained with this wavelength appear to arise primarily from changes in blood volume [7]. For visual stimulation, the contrast of the image was adjusted to 90 cd / m 2 measured at the monitor surface with the background being 2 cd / m 2 . A vertical drifting grating in a spatial frequency of 0.5 cpd and a speed of 28 / s moving rightwards was presented to the animal in a distance of 40 cm (28328). Cortical activity was imaged with a CCD camera (Bischke, CCD 5230p), using a tandem lens (f1.7 / f4.7–5.6, Pentax). The camera recorded pictures at intervals of 1 s, and 100 pictures were taken sequentially, each picture being an average of five frames. In order to cancel the effects of uneven illumination of the cortex and to emphasize responses in the intrinsic signals, the maps were derived by subtracting the image obtained at a particular time from a reference image taken prior to the hundred pictures. For analyzing the data, a cross correlogram was used which is a standard method for estimating the similarity of two waveforms. Cross correlation coefficient at zero lag time (CCF-0) between two waveforms varies between 1 and 21, depending upon the overall phase difference. Waveforms with most of their components being antiphase can have a CCF-0 closer to 21. When two waveforms are identical, the correlation is called autocorrelation, with its coefficient at zero lag time being unit. Autocorrelation is useful to expose the rhythmic components of a waveform. In the present data analysis, the cross correlogram is performed between two waveforms X and Y, according to the follow formula:
O
1 n cor(i) 5 ] X(k) ? Y(i 1 k 2 1) n k 51
Fig. 1. An image of the occipital cortical surface taken with a CCD camera, using green light (540 nm) illumination. Squares indicate the sites at which the intrinsic signals were obtained: sites 1 and 2 were in V1, sites 3 and 4 in V2, and sites 5 and 6 in V4. Inset indicates anterior (A), posterior (P), lateral (L) and medial (M) directions. Scale bar 1 mm.
G.B. Liu et al. / Brain Research 985 (2003) 13–20
where cor(i) is the value of one point on the cross correlogram and n is the length of the waveforms. Animal experiments were approved by the Animal Welfare and Ethnic Committee of the University of Queensland, under a protocol number: PHYS / PH / 347 / 99 / NHMRC.
3. Results Fig. 2A illustrates the spontaneous variations of the intrinsic signals as a function of time over the three cortical areas, V1, V2 and V4. The signals were sampled from six sites, as indicated by the numbered squares in Fig. 1. In order to compare activation at different sites within the same area, two sampling sites were selected in each of the three cortical areas (Fig. 1: 1 and 2 in V1, 3 and 4 in V2, and 5 and 6 in V4). The results illustrated in Fig. 2A show that for each pair of sampling sites, both the
15
waveform and polarity of the signal fluctuations are similar, indicating an intra-area synchronization of spontaneous activation. This was further evaluated by crosscorrelation analysis between the paired sites for each cortical area. Fig. 2B shows the cross-correlograms for the paired sites in V1 (upper curve), V2 (middle curve) and V4 (lower curve), the correlation coefficients (CCF-0) for these three areas are 0.94, 0.72 and 0.86, respectively, indicating a synchrony of spontaneous activation within each individual area. This synchronized activation did not, however, occur between different visual cortical areas. Fig. 2C illustrates the cross-correlograms between V1 and V2 (upper curve), V2 and V4 (middle curve), and V4 and V1 (lower curve): no synchrony peak can be identified at time zero in these cross-correlograms. However, the synchrony between cortical areas changed dramatically when the visual cortex was activated by a visual stimulus. To show this, we presented a small grating (28328, vertically oriented) on the fovea of the monkey’s retina, starting at 30 s after signal sampling and ended at
Fig. 2. Phase relationship of the spontaneous variation of the intrinsic signals within the same cortical area and between different areas. (A) The spontaneous activities sampled at six different sites in V1, V2 and V4. Numbers to the left of each trace refer to the corresponding recording sites illustrated in Fig. 1. (B) Cross-correlation analysis showing the synchrony of spontaneous activity within each area. The curves were computed from the data shown in A. Upper curve, cross-correlogram of the two sites (1 and 2) in V1; middle curve, cross-correlogram between the two sites (3 and 4) in V2; lower curve, cross-correlogram between the two sites (5 and 6) in V4. (C) Cross-correlation analysis showing the asynchrony of spontaneous activity between different visual areas. Upper curve, cross-correlogram of spontaneous activity between V1 (site 1) and V2 (site 3); middle curve, cross-correlogram between V2 (site 3) and V4 (site 5); lower curve, cross-correlogram between V4 (site 5) and V1 (site 1).
16 G.B. Liu et al. / Brain Research 985 (2003) 13–20 Fig. 3. Optical images of the intrinsic signals showing the spatio-temporal responses evoked by continuous visual stimulation. The visual stimulus was a 28328 sinusoidal grating presented at the center of the visual field (to the fovea of the retina), the spatial frequency was 0.5 c / deg, and the speed of drift of the grating was 28 / s. The numbers at the lower right-hand corner of the maps indicate the time (s) at which the individual image was taken. The arrows indicate on and off of the stimulus. Response strength (level of darkness) of the intrinsic signals is represented by the pseudo-colors, as explained at the bottom of the figure. All the maps are normalized with reference to the response maximum. The images were obtained by subtracting the average of the pre-stimulation images from the average of the response images. Seven parallel bands perpendicular to the V1 / V2 border can be seen following this analysis procedure.
G.B. Liu et al. / Brain Research 985 (2003) 13–20
80 s (lasting for 50 s). Fig. 3 shows the spatio-temporal maps evoked by this localized stimulus. About 5 s after the onset of the visual stimulus (the map at 36th second), a dark spot (represent by a yellow pseudo-color in Fig. 3) appeared in the optical image at cortical site where the fovea of the retina is represented [5,14]. The activated area then expanded bilaterally along the border between V1 and V2 [5,14] and enlarged to cover most of the lateral surface of the occipital cortex. The parallel strips manifest the ocular dominance columns due to monocular stimulation. The spread covered an extent of 8 mm parallel to the V1 / V2 border and 4 mm along the ocular dominance columns, with a faster speed (0.46 mm / s) along border than along the ocular dominance columns (0.26 mm / s). The activation faded away 10 s after terminating stimulus presentation (map at the 91st second in Fig. 3). In this condition, a high synchrony of activation was observed between the different cortical areas. The responses of the intrinsic signals to visual stimulation are shown in Fig. 4A. The six curves illustrate the optical responses from V1 (upper pair of curves), V2 (middle pair of curves) and V4 (lower pair of curves), respectively. The two vertical lines indicate the onset and offset of the visual stimuli, respectively. The sampling sites correspond to those in Fig. 1 (1 and 2 from V1, 3 and 4 from V2, and 5 and 6 from V4). It is seen that visual stimulus induced an increase of gray scale in V1 and V2, and an opposite change in V4. The synchrony of the optical responses with the visual stimulus at all the sites sampled is well illustrated in Fig. 4A. The synchrony and phase relationships of the intrinsic signals between the different cortical areas are shown by crosscorrelation analysis (Fig. 4B). The upper curve of Fig. 4B shows the cross-correlograms for V1 (site 1) and V2 (site 4) (CCF-0 coefficient 0.78), the middle curve is the crosscorrelogram for V2 (site 4) and V4 (site 6) (CCF-0
17
20.93), and the lower curve, the cross-correlogram for V1 (site 1) and V4 (site 6) (CCF-0 20.87). The similarity in shape and locations of the peaks in the correlograms suggests that the localized visual stimulation induced a synchronous activation in these three cortical areas, with in-phase activation in V1 and V2 and anti-phase activation in V4. Similar results were obtained from the four hemispheres of two monkeys. When monocular stimulus was used, the cortical activation induced by local stimulation was found to spread along the ocular dominance columns. We have compared the frequency and phase of the oscillatory activities between the activated columns (points 1 and 4 in Fig. 5A) and the non-activated inter-column areas (points 2 and 3, in Fig. 5A). The results (Fig. 5B) showed that the frequency, waveform and phase of the signal oscillations between the activated and non-activated areas were similar, and correlation analysis (Fig. 5C) indicated high synchrony among these sub-areas.
4. Discussion The area of cortex activated by a minimal visual stimulus, referred to as cortical ‘point spread’ (PS) [13], has been investigated in V1 of cat and monkey [6,11]. The fovea is represented near the lower tip of areas 17–18 border, running about 2 mm behind the lunate sulcus on the lateral side of the hemisphere [5,14]. Ocular dominance columns run orthogonal to the 17–18 border where the VM is represented. Since the magnification along the 17–18 border is greater than along the ocular dominance columns, a greater extent and a faster speed of the spread of activation along the 17–18 border than along the ocular dominance columns are expected.
Fig. 4. Responses of the intrinsic signals of three cortical areas to a local stimulation at the fovea. (A) Post-stimulus time histograms of the intrinsic optical signals generated in different cortical areas. The recording sites are indicated in Fig. 1. The data are from the same experiment as shown in Fig. 3. Arrows indicate ON and OFF of the stimulus. (B) Cross-correlation analysis of intrinsic signals between different cortical areas during visual stimulation. The curves were computed from the data shown in A. Upper curve, cross-correlogram of intrinsic signals between V1 (site 1) and V2 (site 4); middle curve, cross-correlogram between V2 (site 4) and V4 (site 6); lower curve, cross-correlogram between V1 (site 3) and V4 (site 6).
18
G.B. Liu et al. / Brain Research 985 (2003) 13–20
Fig. 5. A comparison of the oscillatory activities over the visually activated areas and non-activated areas in V1. (A) Optical images evoked by monocular fovea stimulation. The dark strips show the ocular dominance columns evoked by the stimulus. The squares indicate the sampling sites of the optical signals shown in B. Sites 1 and 4 were from the activated columns, and sites 2 and 3 from the non-activated inter-column areas. (B) The intrinsic signals in response to monocular fovea stimulation recorded from the four sampling sites. Numbers to the left of each trace refer to the corresponding recording sites illustrated in A. (C) Cross-correlation analysis showing the synchrony of the signals between the activated and non-activated areas. The curves were computed from the sampling sites shown in A. Upper curve, cross-correlogram of intrinsic signals between an activated site (site 1) and a non-activated site (site 2), cross-correlation coefficient 0.89; middle curve, cross-correlogram between site 4 (activated) and site 3 (non-activated), cross-correlation coefficient 0.83; lower curve, cross-correlogram between two activated sites (sites 4 and 1), cross-correlation coefficient 0.96.
By a combination of optical imaging and single-unit recording, Das and Gilbert [6] used a 0.58-line segment to compare the activated areas measured with optical recording (optical PS) with those measured with extracellular electrodes (spike PS). They found that the optical PS extended over a 3.2 to 5.2 mm cortical area on the average, and is considerably larger than the spike PS (0.5 by 1 mm in cat’s cortex). This result indicates that 95% of the optical PS area was probably generated by subthreshold activation of the cells within the spike PS. The extent of optical PS observed in the present study is even larger (8 mm along antero-posterial axis and 4 mm along lateromedial axis), possibly due to the use of more effective stimulus (grating). The extent of optical spread is comparable to the extensive integration field of cortical neurons in cat [12] and monkey (Li and Xu, unpublished), and morphologically to the long-range horizontal connections in V1 [8,11]. The latter may underlie the extensive subthreshold activation observed by optical imaging. Another extensive effect evoked by a minimal stimulus was the inter-area synchronization of the low-frequency
oscillation. As is shown in Fig. 2, when no stimulus was presented in the visual field, the spontaneous oscillatory activities (about 0.1 Hz) were synchronized only within each individual area (intra-area synchronization), the synchrony did not, however, occur between different areas. This situation changed when the visual cortex was activated by a localized visual stimulus. In this circumstance, a high synchrony of the spontaneous activities was observed among V1, V2 and V4. As we have recorded only from these three areas, it is not known whether the synchrony also spread to the other visual areas. Bringuier et al. [4] made intracellular recordings in cat area 17, found that a focal visual stimulus evoked in visual cortex a radial wave of activity spreading in the plane of cortical layers at a constant velocity, and that synaptic depolarizing responses to stimuli flashed at increasing distances from the center of the receptive field decreased in strength, whereas their onset latency increased. Girard et al. [9] reported in monkeys, the feedback and feedforward connections between V1 and V2 had similar fast conduction velocities (around 3.5 m / s). In contrast with the rapid conduction of
G.B. Liu et al. / Brain Research 985 (2003) 13–20
feedback and feedforward axons, most of the horizontal fibers appear to be slow conducting indicated by their long latencies. One may assume that visual responses focused at the classical discharge fields are dominated by fast velocity activity, while responses in the unresponsive region surrounding the classical discharge field result from the integration of visual activation waves spread by slowly conducting horizontal axons (|0.1 m / s). The pattern of synchronization observed in the present study might be related to the co-existence of the feedforward visual input (fast velocity) to all three areas studied, and the activity of long-range cortico–cortical connections (slow velocity) intrinsic to a given area. In addition, the current results might imply that the wide-spread synchrony across cortical areas, although induced by visual stimulation, is not a manifestation of visual activation. It is most probably that the synchronization of intrinsic activities might provides a basis for spatial and temporal integration of the visual signals over different cortical areas and among the ocular dominance column. Arieli et al. [1] compared the coherence of the optical signal and the activity of individual neurons recorded from
19
the cat visual areas 17 and 18, with and without the visual stimulation. They found that the spontaneous firing of single neurons was highly correlated with the optical signals and with the local field potential. The coherent activity in both areas was time locked to the spontaneous firing of single neurons over distances exceeding 6 mm. This result indicates that spontaneous activity of single neurons is not an independent process but is time locked to the firing or to the synaptic inputs from numerous neurons. Although the current result that the optical signals showed intra-area synchrony and inter-area asynchrony was somewhat inconsistent with the prediction deducted from the above experiment that neuronal activities obtained between two distant locations may still be coherent, caution needs to be exercised when estimating the correlation between the signals obtained with different methodologies. Brody [3] also found that slow covariations of membrane potentials in tens of seconds generate a conspicuous peak in cross correlograms of the cell pair even though they are not interacting at all. For the purpose of comparison, the result of bi-hemispheric optical recording obtained from area V1 of a cat is
Fig. 6. Spontaneous optical signals recorded from the visual cortex from one adult cat. (A) Schema illustrating the cat brain and the position where the optical signals were recorded. (B) Blood vessel map of the position shown in A. Interhemispheric functional oscillation was estimated by comparing the optical signals from the two square areas. (C) Optical signals recorded from left and right visual cortices shown in B, in a period of 200 s. (D) Autocorrelation functions (upper and middle waveforms) of the optical signals from C, respectively, which demonstrate that the spindle waves are asynchronous between left and right visual cortices. Lower waveform is the cross correlogram from the two signals from C. The low value at the zero lag time (CCF-05 20.2) indicates an interhemispheric asynchrony.
20
G.B. Liu et al. / Brain Research 985 (2003) 13–20
demonstrated here (with the same anaesthetic and experimental procedure as in monkeys). It is apparent that the optical signals obtained from the macaque monkey and cat are different in their morphologies: The signal from the visual cortex of the cat shows more apparent rhythmic components than that from the monkey. The results from the cat indicate that there are two types of rhythmic components appearing on the optical signals. One is a basic rhythmic component with a frequency around 0.2 Hz. When the basic rhythmic components from two visual cortices are compared in terms of their temporal relation, it is observed that they are out of phase or anti-phase for about 65% of the time reckoned. Fig. 6C shows a pair of signals recorded from homologous areas of the left (upper waveform) and right (lower one) areas V1. The lower waveform in panel D is the cross-correlogram between these two signals. The lower value of the CCF-0 indicates that they are asynchronous. The other rhythmic component is the spindle wave, which is a periodic amplitude variation superimposed on the top of the basic component with a period between 10 and 20 s. The spindle wave can be clearly seen in auto-correlation function of each waveform. In Fig. 6, the upper and middle waveforms of panel D are the auto-correlation functions of the signals from panel C, which reveal the spindle waves (arrows). It is apparent that the spindle waves from opposite hemispheres are asynchronous with various durations. Further investigations are needed for the functional significance of the bi-hemispheric asynchrony in terms of the basic components and the spindle wave. However, from the above data, the existence of a functional interhemispheric asynchrony in a wide temporal range is evident. Considering the finding that the spontaneous activities from an individual area are synchronous and from different areas are asynchronous, one may speculate that the asynchrony of spontaneous activity increases with the increase of cortical distance. It’s mechanism, however, remains to be investigated.
Acknowledgements Y.Z., W.F.X. and C.-Y.L. were supported by the Major State Basic Research Program (G2000077800), Natural Science Foundation of China (90208006), Brain and Mind Research Project of CAS (KJCX-07), and Laboratory of Visual Information Processing. G.B.L. and J.D.P. were
supported by a grant from the National Health and Medical Research Council of Australia (Comparative Physiology of Binocular Vision).
References [1] A. Arieli, D. Shoham, R. Hildesheim, A. Grinvald, Coherent spatiotemporal patterns of ongoing activity revealed by real-time optical imaging coupled with single-unit recording in the cat visual cortex, J. Neurophysiol. 73 (1995) 2072–2093. [2] T. Bonhoeffer, A. Grinvald, Optical imaging based on intrinsic signals. The methodology, in: A.W. Toga, J.C. Mazziotta (Eds.), Brain Mapping: The Methods, Academic Press, San Diego, CA, 1996, pp. 55–97. [3] C.D. Brody, Disambiguating different covariation types, Neural Comput. 11 (1999) 1527–1535. [4] V. Bringuier, F. Chavane, L. Glaeser, Y. Fregnac, Horizontal propagation of visual activity in the synaptic integration field of area 17 neurons, Science 283 (1999) 695–699. [5] P.M. Daniel, D. Whitteridge, The representation of the visual field on the cerebral cortex in monkeys, J. Physiol. 159 (1961) 203–222. [6] A. Das, C.D. Gilbert, Long-range horizontal connections and their role in cortical reorganization revealed by optical recording of cat primary visual cortex, Nature 375 (1995) 780–784. [7] R.D. Frostig, E.E. Lieke, D.Y. Ts’o, A. Grinvald, Cortical functional architecture and local coupling between neuronal activity and microcirculation revealed by in vivo, high resolution optical imaging of intrinsic signals, Proc. Natl. Acad. Sci. USA 87 (1990) 6082– 6086. [8] C.D. Gilbert, T.N. Wiesel, Receptive field dynamics in adult primary visual cortex, Nature 356 (1992) 150–152. [9] P. Girard, J.M. Hupe, J. Bullier, Feedforward and feedback connections between areas V1 and V2 of the monkey have similar rapid conduction velocities, J. Neurophysiol. 85 (2001) 1328–1331. [10] A. Grinvald, E. Lieke, R.D. Frostig, C.D. Gilbert, T.N. Wiesel, Functional architecture of cortex revealed by optical imaging of intrinsic signals, Nature 324 (1986) 361–364. [11] A. Grinvald, E.E. Lieke, R.D. Frostig, R. Hildesheim, Cortical point-spread function and long-range lateral interactions revealed by real-time optical imaging of macaque monkey primary visual cortex, J. Neurosci. 14 (1994) 2543–2568. [12] C.Y. Li, W. Li, Extensive integration field beyond the classical receptive field of cat’s striate cortical neurons—classification and tuning properties, Vision Res 34 (1994) 2337–2355. [13] J.T. McIlwain, Visual receptive fields and their images in superior colliculus of the cat, J. Neurophysiol. 38 (1975) 219–230. [14] R.B. Tootell, M.S. Silverman, E. Switkes, R.L. De Valois, Deoxyglucose analysis of retinotopic organization in primate striate cortex, Science 218 (1982) 902–904. [15] D.Y. Ts’o, R.D. Frostig, E.E. Lieke, A. Grinvald, Functional organization of primate visual cortex revealed by high resolution optical imaging, Science 249 (1990) 417–420.
Research report
Spreading and synchronization of intrinsic signals in visual cortex of macaque monkey evoked by a localized visual stimulus Guang Bin Liu a , Ying Zhang b , John Douglas Pettigrew a , Wei Feng Xu b , Chao-Yi Li b , * a
Vision, Touch and Hearing Research Centre, Faculty of Biological and Chemical Sciences, The University of Queensland, Brisbane, Australia b Institute of Neuroscience, Shanghai Institute of Biological Sciences, Chinese Academy of Sciences, Shanghai 200031, China Accepted 22 May 2003
Abstract Spatio-temporal maps of the occipital cortex of macaque monkeys were analyzed using optical imaging of intrinsic signals. The images obtained during localized visual stimulation (IS) were compared with the images obtained on presentation of a blank screen (IB). We first investigated spontaneous variations of the intrinsic signals by analyzing the 100 IBs for each of the three cortical areas. Slow periodical activation was observed in alternation over the cortical areas. Cross-correlation analysis indicated that synchronization of spontaneous activation only took place within each cortical area, but not between them. When a small, drifting grating (28328) was presented on the fovea, a dark spot appeared in the optical image at the cortical representation of this retinal location. It spread bilaterally along the border between V1 and V2, continuing as a number of parallel dark bands covering a large area of the lateral surface of V1. Cross-correlation analysis showed that during visual stimulation the intrinsic signals over all of the three cortical areas were synchronized, with in-phase activation of V1 and V2 and anti-phase activation of V4 and V1 / V2. The significance of these extensive synergistic and antagonistic interactions between different cortical areas is discussed. 2003 Elsevier B.V. All rights reserved. Theme: Sensory systems Topic: Visual cortex: striate Keywords: Optical imaging; Visual cortex; Macaque monkey
1. Introduction The delineation of visual cortical areas by electrophysiological investigations is often problematic due to the long time span needed to map many cortical sites in order to define area boundaries. In addition, the anatomical corroboration of the thus-proposed boundaries comes too late to influence the course of a physiological experiment. Intrinsic signal optical recording, which is based on changes in local blood supply and activity-related metabolism [2,10,15], allows the functional delineation of the boundaries between relatively large areas of visual cortex within the time frame of one recording experiment. It also *Corresponding author. Tel.: 186-21-5492-1777; fax: 186-21-54921778. E-mail address: [email protected] (C.-Y. Li). 0006-8993 / 03 / $ – see front matter 2003 Elsevier B.V. All rights reserved. doi:10.1016 / S0006-8993(03)03049-X
provides a high spatial resolution and reveals the dynamic properties of cortical activation. In the current experiment, we report a novel observation using this technique. In response to a small visual stimulus presented on the fovea of the monkey’s retina, a dark spot appeared in the optical image at the area of the visual cortex corresponding to the retinal location of the stimulus. This area of activation then spread to include the greater part of the occipital cortex. Using cross-correlation analysis, we were able to study the dynamics of this extensive activation across different visual cortical areas. In addition, we observed functional boundaries between recognized visual cortical areas that were based on differences in the timing of spontaneous oscillations in the intrinsic signal recorded optically from different visual cortical areas at the same time. While further work will be needed to define the source of these spontaneous oscillations, the fact that they
14
G.B. Liu et al. / Brain Research 985 (2003) 13–20
are coherent within one area but differ across areas offers a new basis for defining various cortical areas. The present finding also suggests alternative mechanisms to explain how the boundaries are set up in the first place.
2. Methods Data were obtained from the occipital cortex of two anesthetized and paralyzed rhesus monkeys (Macaca mulatta). The methods used for the preparation of the animals were basically those which have been used for previous neurophysiological investigations in our laboratory [12]. Briefly, the monkey was initially anesthetized with ketamine hydrochloride (30 mg per kg body weight). Subsequently, the monkey received an infusion of gallamine triethiodide (10 mg / kg / h), D-tubocurarine chloride (0.25 mg / kg / h), urethane (20 mg / kg / h) and glucose (200 mg / kg / h) in Ringer’s solution (3 ml / h), for muscle relaxation, anesthesia and appropriate physiological condition. End-tidal CO 2 , body temperature and ECG were monitored throughout the experiment and these parameters were maintained within the normal range. Access to the visual cortex was provided by a craniotomy, 22 mm in diameter, so that areas V1, V2, and V4 were exposed (Fig. 1) and the dura was opened. In order to minimize cortical movement due to respiratory and cardiac activity, a stainless steel chamber system was glued to the skull, using dental acrylic. The chamber was filled with silicone oil and closed with a glass window. The exposed cortex was illuminated with light guides, using a 540-nm wavelength interference filter (bandwidth of 30 nm). Signals
obtained with this wavelength appear to arise primarily from changes in blood volume [7]. For visual stimulation, the contrast of the image was adjusted to 90 cd / m 2 measured at the monitor surface with the background being 2 cd / m 2 . A vertical drifting grating in a spatial frequency of 0.5 cpd and a speed of 28 / s moving rightwards was presented to the animal in a distance of 40 cm (28328). Cortical activity was imaged with a CCD camera (Bischke, CCD 5230p), using a tandem lens (f1.7 / f4.7–5.6, Pentax). The camera recorded pictures at intervals of 1 s, and 100 pictures were taken sequentially, each picture being an average of five frames. In order to cancel the effects of uneven illumination of the cortex and to emphasize responses in the intrinsic signals, the maps were derived by subtracting the image obtained at a particular time from a reference image taken prior to the hundred pictures. For analyzing the data, a cross correlogram was used which is a standard method for estimating the similarity of two waveforms. Cross correlation coefficient at zero lag time (CCF-0) between two waveforms varies between 1 and 21, depending upon the overall phase difference. Waveforms with most of their components being antiphase can have a CCF-0 closer to 21. When two waveforms are identical, the correlation is called autocorrelation, with its coefficient at zero lag time being unit. Autocorrelation is useful to expose the rhythmic components of a waveform. In the present data analysis, the cross correlogram is performed between two waveforms X and Y, according to the follow formula:
O
1 n cor(i) 5 ] X(k) ? Y(i 1 k 2 1) n k 51
Fig. 1. An image of the occipital cortical surface taken with a CCD camera, using green light (540 nm) illumination. Squares indicate the sites at which the intrinsic signals were obtained: sites 1 and 2 were in V1, sites 3 and 4 in V2, and sites 5 and 6 in V4. Inset indicates anterior (A), posterior (P), lateral (L) and medial (M) directions. Scale bar 1 mm.
G.B. Liu et al. / Brain Research 985 (2003) 13–20
where cor(i) is the value of one point on the cross correlogram and n is the length of the waveforms. Animal experiments were approved by the Animal Welfare and Ethnic Committee of the University of Queensland, under a protocol number: PHYS / PH / 347 / 99 / NHMRC.
3. Results Fig. 2A illustrates the spontaneous variations of the intrinsic signals as a function of time over the three cortical areas, V1, V2 and V4. The signals were sampled from six sites, as indicated by the numbered squares in Fig. 1. In order to compare activation at different sites within the same area, two sampling sites were selected in each of the three cortical areas (Fig. 1: 1 and 2 in V1, 3 and 4 in V2, and 5 and 6 in V4). The results illustrated in Fig. 2A show that for each pair of sampling sites, both the
15
waveform and polarity of the signal fluctuations are similar, indicating an intra-area synchronization of spontaneous activation. This was further evaluated by crosscorrelation analysis between the paired sites for each cortical area. Fig. 2B shows the cross-correlograms for the paired sites in V1 (upper curve), V2 (middle curve) and V4 (lower curve), the correlation coefficients (CCF-0) for these three areas are 0.94, 0.72 and 0.86, respectively, indicating a synchrony of spontaneous activation within each individual area. This synchronized activation did not, however, occur between different visual cortical areas. Fig. 2C illustrates the cross-correlograms between V1 and V2 (upper curve), V2 and V4 (middle curve), and V4 and V1 (lower curve): no synchrony peak can be identified at time zero in these cross-correlograms. However, the synchrony between cortical areas changed dramatically when the visual cortex was activated by a visual stimulus. To show this, we presented a small grating (28328, vertically oriented) on the fovea of the monkey’s retina, starting at 30 s after signal sampling and ended at
Fig. 2. Phase relationship of the spontaneous variation of the intrinsic signals within the same cortical area and between different areas. (A) The spontaneous activities sampled at six different sites in V1, V2 and V4. Numbers to the left of each trace refer to the corresponding recording sites illustrated in Fig. 1. (B) Cross-correlation analysis showing the synchrony of spontaneous activity within each area. The curves were computed from the data shown in A. Upper curve, cross-correlogram of the two sites (1 and 2) in V1; middle curve, cross-correlogram between the two sites (3 and 4) in V2; lower curve, cross-correlogram between the two sites (5 and 6) in V4. (C) Cross-correlation analysis showing the asynchrony of spontaneous activity between different visual areas. Upper curve, cross-correlogram of spontaneous activity between V1 (site 1) and V2 (site 3); middle curve, cross-correlogram between V2 (site 3) and V4 (site 5); lower curve, cross-correlogram between V4 (site 5) and V1 (site 1).
16 G.B. Liu et al. / Brain Research 985 (2003) 13–20 Fig. 3. Optical images of the intrinsic signals showing the spatio-temporal responses evoked by continuous visual stimulation. The visual stimulus was a 28328 sinusoidal grating presented at the center of the visual field (to the fovea of the retina), the spatial frequency was 0.5 c / deg, and the speed of drift of the grating was 28 / s. The numbers at the lower right-hand corner of the maps indicate the time (s) at which the individual image was taken. The arrows indicate on and off of the stimulus. Response strength (level of darkness) of the intrinsic signals is represented by the pseudo-colors, as explained at the bottom of the figure. All the maps are normalized with reference to the response maximum. The images were obtained by subtracting the average of the pre-stimulation images from the average of the response images. Seven parallel bands perpendicular to the V1 / V2 border can be seen following this analysis procedure.
G.B. Liu et al. / Brain Research 985 (2003) 13–20
80 s (lasting for 50 s). Fig. 3 shows the spatio-temporal maps evoked by this localized stimulus. About 5 s after the onset of the visual stimulus (the map at 36th second), a dark spot (represent by a yellow pseudo-color in Fig. 3) appeared in the optical image at cortical site where the fovea of the retina is represented [5,14]. The activated area then expanded bilaterally along the border between V1 and V2 [5,14] and enlarged to cover most of the lateral surface of the occipital cortex. The parallel strips manifest the ocular dominance columns due to monocular stimulation. The spread covered an extent of 8 mm parallel to the V1 / V2 border and 4 mm along the ocular dominance columns, with a faster speed (0.46 mm / s) along border than along the ocular dominance columns (0.26 mm / s). The activation faded away 10 s after terminating stimulus presentation (map at the 91st second in Fig. 3). In this condition, a high synchrony of activation was observed between the different cortical areas. The responses of the intrinsic signals to visual stimulation are shown in Fig. 4A. The six curves illustrate the optical responses from V1 (upper pair of curves), V2 (middle pair of curves) and V4 (lower pair of curves), respectively. The two vertical lines indicate the onset and offset of the visual stimuli, respectively. The sampling sites correspond to those in Fig. 1 (1 and 2 from V1, 3 and 4 from V2, and 5 and 6 from V4). It is seen that visual stimulus induced an increase of gray scale in V1 and V2, and an opposite change in V4. The synchrony of the optical responses with the visual stimulus at all the sites sampled is well illustrated in Fig. 4A. The synchrony and phase relationships of the intrinsic signals between the different cortical areas are shown by crosscorrelation analysis (Fig. 4B). The upper curve of Fig. 4B shows the cross-correlograms for V1 (site 1) and V2 (site 4) (CCF-0 coefficient 0.78), the middle curve is the crosscorrelogram for V2 (site 4) and V4 (site 6) (CCF-0
17
20.93), and the lower curve, the cross-correlogram for V1 (site 1) and V4 (site 6) (CCF-0 20.87). The similarity in shape and locations of the peaks in the correlograms suggests that the localized visual stimulation induced a synchronous activation in these three cortical areas, with in-phase activation in V1 and V2 and anti-phase activation in V4. Similar results were obtained from the four hemispheres of two monkeys. When monocular stimulus was used, the cortical activation induced by local stimulation was found to spread along the ocular dominance columns. We have compared the frequency and phase of the oscillatory activities between the activated columns (points 1 and 4 in Fig. 5A) and the non-activated inter-column areas (points 2 and 3, in Fig. 5A). The results (Fig. 5B) showed that the frequency, waveform and phase of the signal oscillations between the activated and non-activated areas were similar, and correlation analysis (Fig. 5C) indicated high synchrony among these sub-areas.
4. Discussion The area of cortex activated by a minimal visual stimulus, referred to as cortical ‘point spread’ (PS) [13], has been investigated in V1 of cat and monkey [6,11]. The fovea is represented near the lower tip of areas 17–18 border, running about 2 mm behind the lunate sulcus on the lateral side of the hemisphere [5,14]. Ocular dominance columns run orthogonal to the 17–18 border where the VM is represented. Since the magnification along the 17–18 border is greater than along the ocular dominance columns, a greater extent and a faster speed of the spread of activation along the 17–18 border than along the ocular dominance columns are expected.
Fig. 4. Responses of the intrinsic signals of three cortical areas to a local stimulation at the fovea. (A) Post-stimulus time histograms of the intrinsic optical signals generated in different cortical areas. The recording sites are indicated in Fig. 1. The data are from the same experiment as shown in Fig. 3. Arrows indicate ON and OFF of the stimulus. (B) Cross-correlation analysis of intrinsic signals between different cortical areas during visual stimulation. The curves were computed from the data shown in A. Upper curve, cross-correlogram of intrinsic signals between V1 (site 1) and V2 (site 4); middle curve, cross-correlogram between V2 (site 4) and V4 (site 6); lower curve, cross-correlogram between V1 (site 3) and V4 (site 6).
18
G.B. Liu et al. / Brain Research 985 (2003) 13–20
Fig. 5. A comparison of the oscillatory activities over the visually activated areas and non-activated areas in V1. (A) Optical images evoked by monocular fovea stimulation. The dark strips show the ocular dominance columns evoked by the stimulus. The squares indicate the sampling sites of the optical signals shown in B. Sites 1 and 4 were from the activated columns, and sites 2 and 3 from the non-activated inter-column areas. (B) The intrinsic signals in response to monocular fovea stimulation recorded from the four sampling sites. Numbers to the left of each trace refer to the corresponding recording sites illustrated in A. (C) Cross-correlation analysis showing the synchrony of the signals between the activated and non-activated areas. The curves were computed from the sampling sites shown in A. Upper curve, cross-correlogram of intrinsic signals between an activated site (site 1) and a non-activated site (site 2), cross-correlation coefficient 0.89; middle curve, cross-correlogram between site 4 (activated) and site 3 (non-activated), cross-correlation coefficient 0.83; lower curve, cross-correlogram between two activated sites (sites 4 and 1), cross-correlation coefficient 0.96.
By a combination of optical imaging and single-unit recording, Das and Gilbert [6] used a 0.58-line segment to compare the activated areas measured with optical recording (optical PS) with those measured with extracellular electrodes (spike PS). They found that the optical PS extended over a 3.2 to 5.2 mm cortical area on the average, and is considerably larger than the spike PS (0.5 by 1 mm in cat’s cortex). This result indicates that 95% of the optical PS area was probably generated by subthreshold activation of the cells within the spike PS. The extent of optical PS observed in the present study is even larger (8 mm along antero-posterial axis and 4 mm along lateromedial axis), possibly due to the use of more effective stimulus (grating). The extent of optical spread is comparable to the extensive integration field of cortical neurons in cat [12] and monkey (Li and Xu, unpublished), and morphologically to the long-range horizontal connections in V1 [8,11]. The latter may underlie the extensive subthreshold activation observed by optical imaging. Another extensive effect evoked by a minimal stimulus was the inter-area synchronization of the low-frequency
oscillation. As is shown in Fig. 2, when no stimulus was presented in the visual field, the spontaneous oscillatory activities (about 0.1 Hz) were synchronized only within each individual area (intra-area synchronization), the synchrony did not, however, occur between different areas. This situation changed when the visual cortex was activated by a localized visual stimulus. In this circumstance, a high synchrony of the spontaneous activities was observed among V1, V2 and V4. As we have recorded only from these three areas, it is not known whether the synchrony also spread to the other visual areas. Bringuier et al. [4] made intracellular recordings in cat area 17, found that a focal visual stimulus evoked in visual cortex a radial wave of activity spreading in the plane of cortical layers at a constant velocity, and that synaptic depolarizing responses to stimuli flashed at increasing distances from the center of the receptive field decreased in strength, whereas their onset latency increased. Girard et al. [9] reported in monkeys, the feedback and feedforward connections between V1 and V2 had similar fast conduction velocities (around 3.5 m / s). In contrast with the rapid conduction of
G.B. Liu et al. / Brain Research 985 (2003) 13–20
feedback and feedforward axons, most of the horizontal fibers appear to be slow conducting indicated by their long latencies. One may assume that visual responses focused at the classical discharge fields are dominated by fast velocity activity, while responses in the unresponsive region surrounding the classical discharge field result from the integration of visual activation waves spread by slowly conducting horizontal axons (|0.1 m / s). The pattern of synchronization observed in the present study might be related to the co-existence of the feedforward visual input (fast velocity) to all three areas studied, and the activity of long-range cortico–cortical connections (slow velocity) intrinsic to a given area. In addition, the current results might imply that the wide-spread synchrony across cortical areas, although induced by visual stimulation, is not a manifestation of visual activation. It is most probably that the synchronization of intrinsic activities might provides a basis for spatial and temporal integration of the visual signals over different cortical areas and among the ocular dominance column. Arieli et al. [1] compared the coherence of the optical signal and the activity of individual neurons recorded from
19
the cat visual areas 17 and 18, with and without the visual stimulation. They found that the spontaneous firing of single neurons was highly correlated with the optical signals and with the local field potential. The coherent activity in both areas was time locked to the spontaneous firing of single neurons over distances exceeding 6 mm. This result indicates that spontaneous activity of single neurons is not an independent process but is time locked to the firing or to the synaptic inputs from numerous neurons. Although the current result that the optical signals showed intra-area synchrony and inter-area asynchrony was somewhat inconsistent with the prediction deducted from the above experiment that neuronal activities obtained between two distant locations may still be coherent, caution needs to be exercised when estimating the correlation between the signals obtained with different methodologies. Brody [3] also found that slow covariations of membrane potentials in tens of seconds generate a conspicuous peak in cross correlograms of the cell pair even though they are not interacting at all. For the purpose of comparison, the result of bi-hemispheric optical recording obtained from area V1 of a cat is
Fig. 6. Spontaneous optical signals recorded from the visual cortex from one adult cat. (A) Schema illustrating the cat brain and the position where the optical signals were recorded. (B) Blood vessel map of the position shown in A. Interhemispheric functional oscillation was estimated by comparing the optical signals from the two square areas. (C) Optical signals recorded from left and right visual cortices shown in B, in a period of 200 s. (D) Autocorrelation functions (upper and middle waveforms) of the optical signals from C, respectively, which demonstrate that the spindle waves are asynchronous between left and right visual cortices. Lower waveform is the cross correlogram from the two signals from C. The low value at the zero lag time (CCF-05 20.2) indicates an interhemispheric asynchrony.
20
G.B. Liu et al. / Brain Research 985 (2003) 13–20
demonstrated here (with the same anaesthetic and experimental procedure as in monkeys). It is apparent that the optical signals obtained from the macaque monkey and cat are different in their morphologies: The signal from the visual cortex of the cat shows more apparent rhythmic components than that from the monkey. The results from the cat indicate that there are two types of rhythmic components appearing on the optical signals. One is a basic rhythmic component with a frequency around 0.2 Hz. When the basic rhythmic components from two visual cortices are compared in terms of their temporal relation, it is observed that they are out of phase or anti-phase for about 65% of the time reckoned. Fig. 6C shows a pair of signals recorded from homologous areas of the left (upper waveform) and right (lower one) areas V1. The lower waveform in panel D is the cross-correlogram between these two signals. The lower value of the CCF-0 indicates that they are asynchronous. The other rhythmic component is the spindle wave, which is a periodic amplitude variation superimposed on the top of the basic component with a period between 10 and 20 s. The spindle wave can be clearly seen in auto-correlation function of each waveform. In Fig. 6, the upper and middle waveforms of panel D are the auto-correlation functions of the signals from panel C, which reveal the spindle waves (arrows). It is apparent that the spindle waves from opposite hemispheres are asynchronous with various durations. Further investigations are needed for the functional significance of the bi-hemispheric asynchrony in terms of the basic components and the spindle wave. However, from the above data, the existence of a functional interhemispheric asynchrony in a wide temporal range is evident. Considering the finding that the spontaneous activities from an individual area are synchronous and from different areas are asynchronous, one may speculate that the asynchrony of spontaneous activity increases with the increase of cortical distance. It’s mechanism, however, remains to be investigated.
Acknowledgements Y.Z., W.F.X. and C.-Y.L. were supported by the Major State Basic Research Program (G2000077800), Natural Science Foundation of China (90208006), Brain and Mind Research Project of CAS (KJCX-07), and Laboratory of Visual Information Processing. G.B.L. and J.D.P. were
supported by a grant from the National Health and Medical Research Council of Australia (Comparative Physiology of Binocular Vision).
References [1] A. Arieli, D. Shoham, R. Hildesheim, A. Grinvald, Coherent spatiotemporal patterns of ongoing activity revealed by real-time optical imaging coupled with single-unit recording in the cat visual cortex, J. Neurophysiol. 73 (1995) 2072–2093. [2] T. Bonhoeffer, A. Grinvald, Optical imaging based on intrinsic signals. The methodology, in: A.W. Toga, J.C. Mazziotta (Eds.), Brain Mapping: The Methods, Academic Press, San Diego, CA, 1996, pp. 55–97. [3] C.D. Brody, Disambiguating different covariation types, Neural Comput. 11 (1999) 1527–1535. [4] V. Bringuier, F. Chavane, L. Glaeser, Y. Fregnac, Horizontal propagation of visual activity in the synaptic integration field of area 17 neurons, Science 283 (1999) 695–699. [5] P.M. Daniel, D. Whitteridge, The representation of the visual field on the cerebral cortex in monkeys, J. Physiol. 159 (1961) 203–222. [6] A. Das, C.D. Gilbert, Long-range horizontal connections and their role in cortical reorganization revealed by optical recording of cat primary visual cortex, Nature 375 (1995) 780–784. [7] R.D. Frostig, E.E. Lieke, D.Y. Ts’o, A. Grinvald, Cortical functional architecture and local coupling between neuronal activity and microcirculation revealed by in vivo, high resolution optical imaging of intrinsic signals, Proc. Natl. Acad. Sci. USA 87 (1990) 6082– 6086. [8] C.D. Gilbert, T.N. Wiesel, Receptive field dynamics in adult primary visual cortex, Nature 356 (1992) 150–152. [9] P. Girard, J.M. Hupe, J. Bullier, Feedforward and feedback connections between areas V1 and V2 of the monkey have similar rapid conduction velocities, J. Neurophysiol. 85 (2001) 1328–1331. [10] A. Grinvald, E. Lieke, R.D. Frostig, C.D. Gilbert, T.N. Wiesel, Functional architecture of cortex revealed by optical imaging of intrinsic signals, Nature 324 (1986) 361–364. [11] A. Grinvald, E.E. Lieke, R.D. Frostig, R. Hildesheim, Cortical point-spread function and long-range lateral interactions revealed by real-time optical imaging of macaque monkey primary visual cortex, J. Neurosci. 14 (1994) 2543–2568. [12] C.Y. Li, W. Li, Extensive integration field beyond the classical receptive field of cat’s striate cortical neurons—classification and tuning properties, Vision Res 34 (1994) 2337–2355. [13] J.T. McIlwain, Visual receptive fields and their images in superior colliculus of the cat, J. Neurophysiol. 38 (1975) 219–230. [14] R.B. Tootell, M.S. Silverman, E. Switkes, R.L. De Valois, Deoxyglucose analysis of retinotopic organization in primate striate cortex, Science 218 (1982) 902–904. [15] D.Y. Ts’o, R.D. Frostig, E.E. Lieke, A. Grinvald, Functional organization of primate visual cortex revealed by high resolution optical imaging, Science 249 (1990) 417–420.