---

Complex Spectral Domain ICA, a New Dimension in Biosignal Analysis (S03003.pdf)

Inventors
Jšrn AnemŸller and Scott Makeig

Applications
EEG and Magnetic EEG signal analysis, Signal Separation, ICA
An improved complex signal separation method capable of opening a new window into the dynamics of EEG activity.

Independent component analysis (ICA) is effective in analyzing brain signals and in particular electroencephalographic (EEG) data. However, ICA algorithms presently applied to brain data rely on several idealized assumptions about the underlying processes that may not be fully applicable. Presently ICA analysis of brain data is carried out assuming a linear and instantaneous mixing process that can be expressed mathematically as multiplication by a single mixing matrix. In the standard ICA model, component signal sources are viewed as neural activity occurring in a perfectly synchronized manner within spatially fixed cortical domains. This does not take into account the possible spatio-temporal dynamics underlying neural processes. One way to exhibit more complex dynamics is to assume a convoluting mixing model. Also neglected in the standard ICA model, is the spectral quality of EEG signals. EEG activity has distinctive characteristics in the different frequency bands which may be associated with different physiological processes. These shortcomings can be overcome and EEG signal analysis can be enhanced using the new method of analysis of brain data. The invention is based on spectral decomposition of the sensor signals and subsequent analysis within distinct spectral bands by means of a complex algorithm for independent component analysis. Other recording techniques, such as magnetoencephalogram,(MEG) or functional magnetic resonance imaging (fMRI), and other electrical recordings from the human body such as electromyographic (EMG) and electrocardiographic (ECG) recording can also benefit from the new method..

References
Neural Networks, 16: (2003), 1311-1323

Patent Status:
U.S. Patent Application Published as 2005-0007091

License Terms:
Exclusive Licenses within specified fields of use

Reference_Number: S03003
Contact: Mike White, Ph.D., CLP o Director, OTM o 858.453.4100 x1703 o mwhite@salk.edu

---

Dynamic Signal Processing and Sleep Parametric EEG Automated Recognition System (SPEARS) (S05001.pdf)

Inventors
Philip Low and Terrence Sejnowski

Applications
Sleep Studies, Signal Processing, EEG Analysis
A fast, unsupervised sleep scoring algorithm capable of assessing sleep states from a single EEG channel using known features and a variable number of states.

Currently, sleep studies are most commonly performed using many channels of data from electroencephalograms (EEGs) which are then manually scored, a lengthy, costly and inaccurate procedure. Alternative sleep state determination methods, including artificial neural network classifiers, tend to emulate human performance and are useful for increasing only the speed of determination, not its quality. In addition, EEGs of brain waves tend to have less power at higher frequencies, making it difficult to notice significant trends. As a result, valuable low-power frequency range data is typically disregarded or ignored. This invention provides an improved, fast, unsupervised, and quantitatively rigorous alternative to current EEG signal analysis methods by providing a way to extract hidden information in data (spectral data). The algorithm uses such hidden information to rapidly separate sleep states on a single channel of data. The technologies presented can determine low power frequency range information from spectral data, which can be used in the analysis of a variety of raw signal data. For example, low-power frequency range information within EEG data for a subject from a period of sleep can be used to determine sleep states. Similarly, automated full-frequency spectral EEG signal analysis can be useful for customized analysis including assessing sleep quality, detecting pathological conditions, and determining the effect of medication on sleep states. The speed at which this data analysis can be performed, the customized and unsupervised nature of analysis, and the ability to extract previously unanalyzed low power frequency information make this methodology particularly attractive to a variety of fields of study. The technology can be highly adaptable using a variable number of states, identification rules, adaptable calibration, and variable time and spectral resolution..

References
No publications to date

Patent Status:
U.S. Patent Application filed

License Terms:
Exclusive, Partially Exclusive, Nonexclusive license negotiable

Reference_Number: S05001
Contact: Mike White, Ph.D., CLP o Director, OTM o 858.453.4100 x1703 o mwhite@salk.edu

---

Cutting Device and Methods of Use (S04002.pdf)

Inventors
Michael Neubig

Applications
Laboratory Tools, Medical Devices, Other
A device for cutting through biological tissue, including juvenile rodent skulls, at a controlled depth without disturbing or damaging underlying tissue.

In animal experiments involving surgery, and other applications, it is often desirable to cut through biological tissue without damaging underlying tissue. For example, in certain animal experiments it is necessary to access the brain of a juvenile rodent and for the brain to remain undamaged. Existing methods using tools such as scissors or scalpels require extreme care, and the small distance between the skull and the brain make this a difficult procedure to learn and carry out without damaging the brain. To solve these problems, the present invention relates to a device for cutting biological tissue at a controlled depth to prevent damage to underlying tissue. The cutting blade is adjustable for different applications, providing an improvement over the use of either scissors or scalpels, and a holding member makes the device easier to use. Possible applications include animal or human surgery, or use with materials other than biological tissue. In addition, the device provides a more ergonomic alternative to commonly used skull-cutting devices such as scissors and rongeurs, which with repeated use often cause hand pain or exacerbate conditions such as carpal tunnel syndrome..

References
No publications to date

Patent Status:
U.S. Patent Application filed

License Terms:
Exclusive, Partially Exclusive, Nonexclusive license negotiable

Reference_Number: S04002
Contact: Mike White, Ph.D., CLP o Director, OTM o 858.453.4100 x1703 o mwhite@salk.edu

---

Surface Segmentation from Luminance and Color Differences (S03008.pdf)

Inventors
Donald MacLeod, Geoffrey Boynton and Ione Fine

Applications
Automatic Segmentation of Targets in Satellite Photos, Segmenting Different Tissue Types in Medical Imaging, Use of Color to Segment and Identify "Meaningful Regions" within a Scene.
A Bayesian model used to automatically segment scenes based on the luminance and color statistics of a particular image set.

The invention describes a plausible segmentation model based on luminance and color differences within natural scenes. Color and luminance differences show a striking lack of independence. When two points fall on the same surface, differences in luminance and color between the two points tend to be small. Conversely, when two points fall on different surfaces, the distribution of luminance and color differences tend to be larger. These statistics for color and luminance differences are not easily captured by correlation statistics or independent component analysis. However the Bayesian model of the invention captures these dependencies well. The model (with no free parameters) based on color and luminance pairwise differences, segments images similarly to human observers, both with an image set based on natural scenes and with a very different novel image set containing both natural and man-made environments. The model is based on the statistics governing differences in luminance and color between neighboring pixels in a given scene (such as a picture of a natural scene, a mammogram image or a satellite photo). For pixels of a given separation, differences in luminance and color are not independent, a change in luminance predicts a change in color and vice versa. Differences in luminance and color between nearby pixels can be modeled by assuming that the probability of belonging to the same surface decreases exponentially with the distance between the two pixels. The Bayesian model based on these statistics can be used to automatically segment scenes based on the luminance and color statistics of the particular image set..

References
J. Opt. Soc. Am. A., 20, No. 7, July 2003

Patent Status:
U.S. Patent Application published as 2005-0058351

License Terms:
Non-exclusive and Exclusive by Field of Use Licenses Negotiable

Reference_Number: S03008
Contact: Dave Odelson, Ph.D. o Senior Licensing Executive o 858.453.4100 x1223 o dodelson@salk.edu

---

An Adaptive System for Broadband Signal Discrimination In A Channel with Reverberation (S93009.pdf)

Inventors
Terrence Sejnowski and Shaolin Li

Applications
Sound separation in noisy environments, Telecommunications, Underwater acoustic telemetry
System for discriminating among multiple signals to recover information

Sound waves in the air and underwater often travel over several paths and interfere with each other, making it difficult to transmit information over a long distance. There is a need for reliable and useful methods to rapidly separate these signals and improve transmissions. This invention relates to a system for separating mixed signal sources by processing a received signal through the combination of a beamforming network and an adaptive Herault-Jutten (HJ) network. The conventional Herault-Jutten network is useful for separating independent signals that have been linearly mixed, but cannot separate a mixture of several independent signals in free field conditions because of the propagation time delays between sources and sensors. The system of this invention combines planar beamforming techniques with a conventional HJ network to adaptively distinguish among signals having delays introduced by the propagation medium. A new sensor filter scheme is introduced to eliminate beamforming variation with frequency over the band of interest. The resulting system could be used in underwater acoustic telemetry for retrieval of data from instruments on the sea floor, in cellular telephone communication to increase channel capacity of telecommunications radio links, and in teleconferencing microphones for separating voices in noisy rooms with poor acoustics. This system could be implemented in an inexpensive VLSI chip and manufactured inexpensively..

References
IEEE Journal of Oceanic Engineering 20: 73-79 (1995)

Patent Status:

---

Miniature Acousto-Optic 3D Scanning Microscope (S07015.pdf)

Inventors
Dejan Vucinic, Terrence Sejnowski

Applications
Cancer Imaging, Brain Research, Microsurgery, Surgical Aid, Clinical Research
Miniature, lightweight wearable multiphoton microscope with microsecond 3D random access capable of deep tissue imaging

Multiphoton imaging is one of the most powerful and most actively developed techniques for imaging of brain activity and morphology, and is rapidly finding new uses in imaging of living vasculature, lymph nodes, as well as detection of cancerous cells in blood vessels through intact skin. Existing commercial imaging systems are, however, very large and expensive (over $400,000), restricting experiments in vivo to preparations that can be tailored to the microscope. Even existing handheld confocal microscopes cost in excess of $40,000 and weigh two pounds or more. Researchers at the Salk Institute have recently invented a means of shrinking a complete acousto-optic scanning microscope into a package under 3" in size and weighing as little as 20 grams (less than _ of an ounce), suitable for head-attached or handheld use. Being compatible with multiphoton excitation, it also permits imaging deep within scattering tissue. Incorporating a new system of beam steering using acousto-optic deflectors, this microscope offers many advantages over the more common galvanometer-based and piezoelectric systems, including: o Three-dimensional scanning with no moving parts; o Inertia-less 3D random access in tens of microseconds, two orders of magnitude faster than most mechanical scanning systems; o Practically perfect repositioning accuracy when driven by digital signals; and o Beam intensity modulation built into the physical mechanism of deflection. In addition, the Salk method makes it possible to build a small and fast microscope that can be attached to a moving experimental subject, enabling a far richer set of experiments to be carried out than possible with traditional stationary microscopes. Having no moving parts, the Salk design can be made more rugged and less sensitive to movement and vibration than mirror-based scanning microscopes, making it extremely well suited for diagnostic and treatment uses in a clinical setting, e.g. as a surgical aid in non-invasive microsurgery of skin conditions or vasculature. The device could also be used as a replacement for handheld confocal microscopes for imaging cancer, capable of obtaining higher resolution images deeper into the skin, or for live imaging of melanin in skin tissue, or blood flow in small vessels in and around tumors..

References
None to date

Patent Status:
U.S. Patent Application Filed 1/17/2008

License Terms:
Exclusive or Non-Exclusive Licenses available by Field of Use

Reference_Number: S07015
Contact: Dave Odelson, Ph.D. o Senior Licensing Executive o 858.453.4100 x1223 o dodelson@salk.edu