Optimized birefringence changes during isolated ... - OSA Publishing

0 downloads 0 Views 345KB Size Report
Apr 10, 2005 - Single trial, birefringence signals associated with action potentials from ... narrowband filtered 100-W halogen light, and birefringence signals ...
Optimized birefringence changes during isolated nerve activation Amanda J. Foust, Roxana M. Beiu, and David M. Rector

Single trial, birefringence signals associated with action potentials from isolated lobster nerves were optimized with high-intensity light-emitting diodes (LEDs) and glass polarizers. The narrow spectral output of the LEDs allowed us to select specific wavelengths, increasing the effectiveness of the polarizers and minimizing the stray light in the system. The LEDs produced intensity profiles equivalent to narrowband filtered 100-W halogen light, and birefringence signals were comparable or superior in size and clarity to halogen lamp recordings. The results support a direct correlation between signal size and polarizer extinction coefficient. Increasing the sensitivity of birefringence detection through the use of LED light sources could ameliorate noninvasive brain imaging techniques that employ fast optical consequences associated with action potential propagation. © 2005 Optical Society of America OCIS codes: 170.3660, 000.1430.

1. Introduction

Each year, neuroscience experiences exponential growth investigating neural function with highdensity electrical recordings. However, neural network and circuitry investigations increasingly require noninvasive functional brain-imaging techniques with higher temporal and spatial resolution than can be achieved with electrical measurements. More desirable methods would provide information on the activity of single neurons over large populations without damaging tissue.1 However, current technical limitations are difficult to overcome. For example, fast optical signals associated with neuronal activity have been recorded noninvasively in humans,2,3 but the spatial resolution is poor (⬃1 cm) and requires many (100 –1000) averages for a good signal-to-noise ratio (SNR). These techniques pass light through the skin and skull and measure changes in scattering and absorption by the neural tissue. Scattering changes are typically caused by neural swelling and other mechanisms that reduce scatter-

The authors are with the Department of Veterinary and Comparative Anatomy, Pharmacology, and Physiology, Washington State University, 405 Wegner Hall, Pullman, Washington 99164. M. Rector’s e-mail address is [email protected]. Received 9 July 2004; revised manuscript received 26 October 2004; accepted 5 December 2004. 0003-6935/05/112008-05$15.00/0 © 2005 Optical Society of America 2008

APPLIED OPTICS 兾 Vol. 44, No. 11 兾 10 April 2005

ing particles within the cell, such as vesicle fusion and membrane unfolding.1 Scattering signals, however, are small, and they appear on a significant amount of background light. Any method that can improve the contrast of optical changes within the tissue will enhance the size of these signals relative to background and noise. Birefringence and 90° scattering signals from in vitro nerve bundle preparations show that light intensity changes associated with birefringence are at least an order of magnitude greater than scattering.4,5 Additionally, the birefringence change appears 2–3 ms earlier and exhibits more temporal structure than the 90° scattered light change.5 Increasing the SNR by employing birefringence related intensity changes may enhance in vivo applications of this technique. Several molecular events in the neuronal membrane produce birefringence changes during the propagation of an action potential. For example, linearly polarized light incident on a resting anisotropic nerve bundle becomes elliptically polarized from components such as phospholipids and membrane proteins. As sodium enters the cell during the initial depolarization, water follows the sodium because the sodium ion tends to be more hydrated than potassium. Water movement causes swelling in the axon, and results in a small shift of the nerve’s refractive index. Additionally, peptide bonds within voltagegated sodium channels reorient in response to membrane depolarization, resulting in a change in birefringence that occurs instantaneously with the activation of the channels.6,7 Some evidence that mi-

crotubules could be involved in the birefringence signal also exists.8 Therefore, detection of changes in birefringence offers great possibilities for achieving high-temporal resolution in locating neural activity. However, these fast optical responses are small and easily obscured by noise.9 The purpose of these experiments lies in improving the feasibility and efficacy of making in vitro birefringence measurements, ideally in single-pass trials. From their infancy, birefringence recordings from in vitro nerves were made by use of halogen lamps.4 The halogen light source was advantageous in terms of stability and high brightness, but it also emitted long-wavelength light (⬎800 nm) that was difficult to exclude with polarizers or filters, resulting in a significant amount of background light and poor extinction coefficients. Recent developments have produced LEDs with high-intensity light emission at low power and cost, eliminating the necessity for high-current power supplies and heat filters. In fact, LED intensity is now comparable with bandpass filtered light from a 100-W halogen source. Green LEDs (520–560 nm) and red LEDs ( 660–700 nm) are readily available and compatible with standard plastic or glass polarizers. Infrared LEDs require special polarizers but have advantages for depth penetration in noninvasive studies.10 The ability of two crossed polarizers to exclude light in the absence of a birefringent material should determine the strength and clarity of measured birefringence signals. 2. Methods

Birefringence recordings were made with nerves removed from lobster legs, Homarus americanus. Each nerve was extracted by use of the Furwasawa pulling-out method and then tied with sutures at both ends to prevent the cytosol leakage.11 The size of the extracted nerve depended on its position on the body of the lobster. The largest nerves came from the legs closest to the claws (denoted as “first walking legs”) and the smallest from the legs closest to the tail (denoted “fourth walking legs”). Nerve size ranged between 30 and 60 mm in length and 0.5 and 2 mm in diameter. We placed each isolated nerve in a recording chamber flooded with lobster Ringers solution. The chamber featured a central, rectangular well with four smaller wells on either side (Fig. 1). Each well was fitted with a silver electrode and isolated with Vaseline. Two electrodes on one side of the large well delivered the current pulse stimuli, and two electrodes on the other side of the well recorded the resulting electrical response. The bottom of the central well contained a narrow hole covered with a microscope slide through which the light was focused during optical recordings.5 Three light sources were used in turn: a red LED (665 nm, 2800 mcd, 5-mm diameter; Panasonic Model LN261CAL), a green LED (520 nm, 2400 mcd, 5-mm diameter; Stanley Electric Model DG5306X), and an IR LED [880 nm, 24 mW, 5-mm⬎ diameter; Roithner Lasertechnik (Germany) Model ELD-880-525]. Figure 2

Fig. 1. Experimental setup. LED light (a) passed through a polarizer (d) [after optional collimation through a f ⫽ 27 mm condenser lens (b) and aperture (c) when the IR LED was used], and was projected through a window (j) onto the nerve bundle within the chamber (e) with the long axis of the nerve oriented at 45° with respect to the E-vector. The transmitted, birefringent light passed through a second polarizer (f) with axis of polarization orthogonal to that of the first polarizer and was detected by a photodiode (g). Silver electrodes (h) located in each small well (i) in the chamber were used to stimulate and record from the nerve.

shows emittance curves for each LED, along with the broader band, filtered halogen light. The LEDs were driven by a dc laboratory power supply (LH 18-20, Sorensen, San Diego, California) or 3V battery. The green and red LEDs exhibited a 15° or less output dispersion. The light passed through a polarizer positioned at 45° with respect to the long axis of the

Fig. 2. Emittance curves for the red (665-nm, RED), green (520nm, GRN), and IR (880-nm, IR) LEDs and filtered halogen lamp (HAL) plotted on a log scale. 10 April 2005 兾 Vol. 44, No. 11 兾 APPLIED OPTICS

2009

Fig. 3. Transmittance curves for the three sets of crossed polarizers: (a) plastic, (b) glass, (c) IR glass. The IR glass polarizers excluded the most light, yielding high ECs and larger amplitude signals than either the regular glass or the plastic polarizers but was not efficient for visible light.

nerve. The polarized light then passed through the nerve and another polarizer that was oriented at 90° with respect to the first polarizer. A photodiode (UDT-555UV兾LN, UDT Sensors, Hawthorne, California) detected the change in transmitted light intensity, and the signal was digitized and recorded with 1000⫻ amplification. Since the broad dispersion of the IR light created significant reflections and background illumination within the optical setup, we utilized an aperture and collimator lens to reduce IR reflections internal to the apparatus. Effective polarizer exclusion of light not altered by the nerve was quantified by calculation of an extinction coefficient (EC) (EC ⫽ max兾min) in the absence of the nerve. Max refers to light intensity measured with the axes of the polarizers parallel (uncrossed), and min refers to the light measured with the polarization axes perpendicular (crossed). We typically used extinction coefficients between 300 and 2000. Percent transmittance curves for three sets of crossed polarizers are shown in Fig. 3. The glass IR polarizers (Edmund Industrial Optics, Barrington, New Jersey, Fig. 3c) transmitted the least light when crossed, resulting in high ECs. Unfortunately, the IR polarizers were not as effective at excluding visible light, so we used a visible polarizer set (Linos Photonics, Fig. 3b) for the red and the green LEDs. We predict that we will achieve larger signals with better polarizers compared with the inexpensive plastic polarizer set (Edmund Industrial Optics, Barrington, New Jersey, Fig. 3a). We seated the extracted nerve in the recording chamber and stimulated with short current pulses at random intervals between 1 and 2 s, each lasting 0.2 ms. Stimulus intensity ranged between 0.1 and 3.0 mA, supplied by dc isolated stimulator (Model A365R, World Precision Instruments, Inc., Sarasota, Florida). A healthy nerve produced a single action 2010

APPLIED OPTICS 兾 Vol. 44, No. 11 兾 10 April 2005

Fig. 4. Signal size is directly proportional to polarizer EC. The top four traces show birefringence signals obtained with a green LED that begins with the maximum EC and is decreased 25% in each subsequent trace by rotation of the analyzing polarizer. Signals contain 10 averages. The bottom trace displays the electrical physiological (EPH) response, and the vertical line indicates the time of stimulus.

potential with stimulus strength under 0.2 mA. After we assessed the nerve health, we began stimulation, and recorded the transmitted light intensity with the electrical response. We calculated the birefringence response as a ratio of the change in light intensity measured during an action potential to the light intensity measured at resting state (dI兾I). Since the birefringence signal always decreases with activation, we have inverted the traces so they correspond more closely with the population action potential. However, all numbers presented have negative values. 3. Results

LED spectral specificity increased the polarizer extinction coefficients, resulting in larger signals. The dI兾I signals exhibit a decrease in maximum amplitude proportional to the decrease in extinction coefficient (Fig. 4). In comparing the optical signal directly to the electrical recording of the population action potentials, we noticed several important differences. First, there is a much slower recovery of the birefringence signal, probably because it may take longer for water to leave the cell after the initial swelling. If voltage-sensitive sodium channels are involved, then the birefringence signal might not recover until the channels returned to their reactivated configuration. Figure 5 illustrates the correlation between signal size and extinction ratio of the polarizers and supports the hypothesis that better exclusion of nonspecific light results in a stronger birefringence signal. The LEDs radiated stable, spectrally specific light that elicited high optical responses both in amplitude and in SNR. The maximum extinction coefficient achieved with the red and green LEDs was 2100. Single-pass trials (Fig. 6) exhibit intensity changes

-2 x 10

-4

GRN

10 ms

RED

Fig. 5. Linear relation between max signal amplitude (dI兾I) and %EC. A regression of %EC with %dI兾I amplitude shows a slope of 101 and R2 ⫽ 0.997.

greater than ⫺2 ⫻ 10⫺4dI兾I with 11:1 SNR. Averaging additional trials increased the SNRs to 30:1 in both cases. The birefringence signals evoked with the IR LED were larger than either the green or the red. Figure 6 displays the greatest dI兾I result (⫺4 ⫻ 10⫺4) for the IR LED collected in only five averages with a 56:1 SNR. Since most of the noise observed in the birefringence signal comes from shot-noise sources (random photon events), the recorded noise amplitude will always be proportional to the light intensity. If more of the measured light contains signal because of better exclusion of background, nonspecific light, then the noise will appear reduced relative to the signal.

IR

HAL

EPH

Fig. 6. Gray traces show single-pass signals obtained with the green (GRN) and red (RED) LEDs with SNRs of 11:1. Black lines overlaying the gray display 15 average traces with 30:1 SNRs. The gray IR trace (IR) contains five averages with 56:1 SNR, increasing to 127:1 (black line) after 13 averages. The gray halogen trace (HAL) shows a single-pass trial, increasing to 20 averages in the black trace.

4. Discussion

Each LED source yielded birefringence recordings of comparable or superior size and clarity to optical measurements made with halogen lamps. Previously, birefringence experiments conducted with the same polarizers with halogen light sources and short-pass heat filters were accompanied by extinction coefficients of 400, since the edges of the filtered halogen emission profile lie outside the excluded wavelengths in the crossed polarizer transmittance curve (Figs. 2 and 3). The spectral specificity of the LEDs allows fine control over the light source wavelength, ameliorating our ability to optimize the polarizers, and resulting in large optical responses with high SNR. LED extinction coefficients were above 2000, thus we would expect a fivefold increase in extinction coefficient to result in a fivefold increase in signal amplitude. The signals obtained with the IR LED light source (peak emission wavelength 800 mm) produced optical signals that were often twice as large as those obtained with the green or the red LEDs. This dramatic improvement may occur simply because the extinc-

tion coefficients of the IR polarizers are superior to the visible polarizers. However, there is another more intriguing possibility. Since index of refraction should decrease with increasing wavelength for most materials, the retardance of a given material might be wavelength dependent.12 While studies conducted by Tasaki et al.13 maintain that no significant birefringence and scattering wavelength dependence arises in the 400 – 650-nm range, our results support an improvement in the dI兾I signal at 880 mm over shorter wavelengths, requiring additional study and exemplifying the advantages of IR light in noninvasive applications. Birefringence measurements from an intact mammalian cortex poses significant challenges such as the nonuniform orientation of nerve fibers characteristic of in vivo tissue. Additionally, these studies were performed on nonmyelinated, invertebrate neurons, and we anticipate that scattering properties of myelin sheathes could confound attempts to detect birefringence in vertebrate neural tissue. In vivo application 10 April 2005 兾 Vol. 44, No. 11 兾 APPLIED OPTICS

2011

of this technique also requires making these measurements with reflected rather than transmitted light. Currently, to address in vivo applications, we are further investigating birefringence measurement in reflectance mode with a polarizing beam splitter. Birefringence offers immense opportunities to investigate neural function and mechanisms on several levels. For decades, scientists have been using this optical property to probe the cellular and the molecular nature of action potential propagation. The technical advancements in light sources and polarizers applied in this study have made the birefringence signal detection system less complex and less expensive while increasing signal amplitude and SNR. Such optimizations coupled with in vivo application may lead to the development of a noninvasive technique for monitoring neuron activity patterns with high spatial and temporal resolution. References 1. D. W. Hochman, “ Intrinsic optical changes in neuronal tissue,” Neurosurg. Clin. N. Am. 8, 393– 412 (1997). 2. J. Steinbrink, M. Kohl, H. Obrig, G. Curio, F. Syre, F. Thomas, H. Wabnitz, H. Rinneberg, and A. Villringer, “ Somatosensory evoked fast optical intensity changes detected non-invasively in the adult human head,” Neurosci. Lett. 291, 105–108 (2000). 3. G. Gratton, M. Fabiani, P. M. Corballis, and E. Gratton, “Noninvasive detection of fast siganls from the cortex using

2012

APPLIED OPTICS 兾 Vol. 44, No. 11 兾 10 April 2005

4.

5.

6. 7.

8. 9.

10.

11. 12. 13.

frequency-domain optical methods,” Ann. N. Y. Acad. Sci. 820, 286 –298 (1997). L. B. Cohen, R. D. Keynes, and B. Hille, “ Light scattering and birefringence changes during nerve activity,” Nature 218, 438 – 441 (1968). K. M. Carter, J. S. George, and D. M. Rector, “Simultaneous birefringence and scattered light measurements reveal anatomical features in isolated crustacean nerve,” J. Neurosci. Methods 135, 9 –16 (2004). D. Landowne, “Measuring nerve excitation with polarized light,” Jpn. J. Physiol. 43, S7–S11 (1993). D. Landowne, “Molecular motion underlying activation and inactivation of sodium channels in squid giant axons,” J. Membrane Biol. 88, 173–185 (1985). R. Oldenbourg, E. D. Salmon, and P. T. Tran, “ Birefringence of single and bundled microtubules,” Biophys. J. 7, 645– 654. R. A. Stepnoski, A. LaPorta, F. Raccuia-Behling, G. E. Blonder, R. E. Slusher, and D. Kleinfeld, “Noninvasive detection of changes in membrane potential in cultured neurons by light scattering,” Proc. Natl. Acad. Sci. 88, 9382–9386 (1991). H. R. Eggert and V. Blazek, “Optical properties of human brain tissue, meninges, and brain tumors in the spectral range of 200 to 900 nm,” Neurosurgery 21, 459 – 464 (1987). K. Furusawa, “The depolarization of crustacean nerve by stimulation or oxygen want,” J. Physiol. A 67, 325–342 (1929). H. D. Young and R. A. Freedman, University Physics (AddisonWesley, Reading, Mass., 1996). I. Tasaki, A. Watanabe, R. Sandlin, and L. Carnay, “Changes in fluorescence, turbidity, and birefringence associated with nerve excitation,” Proc. Natl. Acad. Sci. 61, 883– 888 (1968).