Optical Imaging of Intrinsic Signals with Blue Light

Transcription

1 Optical Imaging of Intrinsic Signals with Blue Light Andrei Cimponeriu and Ehud Kaplan The Mount Sinai School of Medicine, New York, NY Phone: (212) Poster NE at SPIE's Opto Northeast and Imaging 2001, April 2001, Rochester, NY

2 Introduction Optical Imaging of Intrinsic Signals (OIS; reviews in [1]-[2]) uses signals that are intrinsic to the exposed cortical tissue to visualize neural activity. The tiny local changes in reflected light are mainly caused by changes in absorption of deoxyhemoglobin (Hb) and of oxyhemoglobin (HbO 2 ). OIS is traditionally performed using red (620nm) or orange (605nm) light. Because the changes in OIS are very small, on the order of , noise is important and the signal is difficult to distinguish from noise. Absorption spectra, in particular the molar extinction coefficients (MECs) of Hb and HbO 2 at these wavelengths, are known ([3]-[4] and the references therein). At the above wavelengths, Hb absorption is dominant, with MEC of and cm -1 /M, respectively. However, in blue and violet light the Hb and HbO 2 MECs exceed cm -1 /M. This suggests that a significant increase in the signal can be obtained by using these wavelengths.

3 Hemoglobin absorption spectra Estimates of Hemoglobin absorption spectra by Scott Prahl 10 6 HbO 2 Hb Molar Extinction Coefficient (cm -1 /M) Wavelength (nm) Choosing the wavelengths: 12 Absorption coefficient ratios HbO 2 /Hb 10 Hb/HbO Wavelength (nm)

4 Experimental Setup : Optical Imaging of Intrinsic Signals

5 Experimental Details Preparation : Anesthetised, paralyzed cat. Exposed primary visual cortex. Visual stimuli : Two conditions: horizontal and vertical bars. Stimuli presented from frame 10 to 110. Ten loops including both conditions. Optical imaging : Photometrix PXL CCD camera Spatial scale: 45µm square pixels (after 3x3 binning) 51 x 81 pixels area shown ( 2.3mm x 3.6mm), cropped from 96 x 128 pixels frames. Cortical illumination: 630 nm, 430 nm, 466 nm LEDs Exposure time chosen such that the CCD wells get approximately filled. Resulting temporal scales: 630 nm : 200 ms/frame 430 nm : 800 ms/frame 466 nm : 450 ms/frame LEDs spectral radiance : Radiance, linear scale LEDs Spectral Radiance 630nm 430nm 466nm Wavelength [nm]

6 Results: 630 nm reflected signal and its time course

7 Results: 630 nm differential map and two sections

8 Results: 630 nm shot noise level and signal to noise ratio Shot noise level determination based on Singular Value Decomposition [5], [6] Peak-to-peak signal level determination based on differential average run shot noise level, 630nm\b_00_00.00a.b experimental data simulated noise σ = eigenvalue, log scale index Noise: x = 2 128, y = 2 96 (camera non-uniformities discarded), single run. σ shot noise = 15.9 Signal: Sig p-p = 42.4 Signal-to-noise ratio: SNR = Sig p-p / σ shot noise = 2.67

9 Results: 430 nm reflected signal and its time course

10 Results: 430 nm differential map and two sections

11 Results: 430 nm shot noise level and signal to noise ratio Shot noise level determination based on Singular Value Decomposition [5], [6] Peak-to-peak signal level determination based on differential average run shot noise level, 430nm\b_00_00.00a.b experimental data simulated noise σ = eigenvalue, log scale Noise: x = 2 128, y = 2 96 (camera non-uniformities discarded), single run. σ shot noise = 12.6 Signal: Sig p-p = 209 index Signal-to-noise ratio: SNR = Sig p-p / σ shot noise = 16.6

12 Results: 466 nm reflected signal and its time course

13 Results: 466 nm differential map and two sections

14 Results: 466 nm shot noise level and signal to noise ratio Shot noise level determination based on Singular Value Decomposition [5], [6] Peak-to-peak signal level determination based on differential average run shot noise level, 466nm\b_00_00.00a.b experimental data simulated noise σ = eigenvalue, log scale Noise: x = 2 128, y = 2 96 (camera non-uniformities discarded), single run. σ shot noise = 13 Signal: Sig p-p = 185 index Signal-to-noise ratio: SNR = Sig p-p / σ shot noise = 14.2

15 Comparing the differential maps at the three λs The differential maps obtained at 630, 430, 466 nm. Pairs of contour plots : nm, nm, nm.

16 Convolving the 630 nm map with a Gaussian The 630 nm differential map convolved with a Gaussian(σ). σ = 3, 4, 5 pixels (135, 180, 225 µm) Contour plots of the convolved 630 nm map and of the 430 nm map.

17 Conclusions Vertical and horizontal sections through the three differential maps. The differential signal is about 4.9, respectively 4.4 times larger at 430 nm and 466 nm than at 630 nm. The signal to shot noise ratio is about 6.2, respectively 5.3 times larger at 430 nm and 466 nm than at 630 nm. The 430 nm and 466 nm differential maps are very similar. They are close to blurred versions of the 630 nm map.

18 References [1] T. Bonhoeffer and A. Grinvald, "Optical imaging based on intrinsic signals", in A. W. Toga and J. C. Mazziotta (Eds.), Brain Mapping. The Methods. New York: Academic Press, pp , [2] A. Grinvald, D. Shoham, A. Shmuel, D. Glaser, I. Vanzetta, E. Shtoyerman, H. Slovin, C. Wijnbergen, R. Hildersheim and A. Arieli, "In-vivo optical imaging of cortical architecture and dynamics", in U. Windhorst and H. Johansson (Eds.), Modern Techniques in neuroscience research. New York: Springer, pp , [3] V. Tuchin, Tissue Optics, Light scattering methods and instruments for medical diagnosis. SPIE Press, [4] Scott Prahl, Optical Absorption of Hemoglobin, Oregon Medical Laser Center Webpage, < /spectra/hemoglobin/index.html> [5] A. Cimponeriu, Noise analysis of our Photometrics PXL camera. Technical report, Laboratory of Visual Neuroscience and Applied Mathematics, Mt. Sinai School of Medicine, New York, NY, August 14-30, [6] A. Cimponeriu, Optical imaging using blue light. Technical report, Laboratory of Visual Neuroscience and Applied Mathematics, Mt. Sinai School of Medicine, New York, NY, January 10, 2001.