Mary Lou Jepsen Launches Startup to Develop Wearable MRI
In 2016, Mary Lou Jepsen launched Openwater, which is a startup intent on developing devices that measure brain activity, like fMRI. Its aim is to make non-invasive portable devices which cost roughly the same as a smartphone. These devices will measure brain activity with similar or improved resolution to conventional MRI machines without requiring the strong magnetic field, helium, or the large expenses for use and maintenance.
Her purpose is to innovate the medical and health fields and to realize telepathy within the next 8 years.
Here, we are going to predict what sort of device Jepsen and her team are developing and the technologies that will be implemented. Our predictions will be based on the speech given by Jepsen at the 2016 i2i Workshop at the NYU School of Medicine where she presented the research basis for Openwater. Some of the technologies described during the workshop already exist and have successfully been applied to humans, while others are at animal stages of testing or still require basic verification through physics and algorithms.
While Jepsen claims telepathy will be available within the next 8 years, we expect Openwater to release two installments of their device: the first, which will be developed in the immediate future, using existing technologies and a later model that will include technologies currently being investigated and verified through physics and mathematics.
What Will the First Product Entail?
The research coming out of Washington University, which Jepsen introduced, is based on a technology called high-density diffuse optical tomography (H-D DOT).
The Principles and Abilities of H-D DOT Which Will Likely be Used as the Basic Technology for Openwater.
By measuring near-infrared light, near-infrared spectroscopy (NIRS) can be used to measure cerebral blood flow and the concentrations of both oxygenated and deoxygenated hemoglobin.
Near-infrared light (wavelength 700~900 nm) transmits through biological tissues like brain matter. Hemoglobin in blood and cytochrome c oxidase found inside mitochondria absorb near-infrared light. When neural cells are activated, the consumption of oxygen and glucose increases, resulting in a change in cerebral blood flow and the concentration of oxygenated and deoxygenated hemoglobin (Hb). Using these basic principles allows NIRS and DOT to predict the state of neural activity. This is achieved by observing the changes in light received a few centimeters away from the brain where near-infrared light is irradiated.
Generally, NIRS uses continuous wave light to predict changes in concentration based on the Beer-Lambert law. Unfortunately, NIRS alone is insufficient as it cannot quantitatively calculate the change in Hb concentration linked to brain activity; furthermore, NIRS readings are often confounded by blood mobilization in the skin.
The introduction of DOT is believed to compensate for these shortcomings. Generally, DOT uses picosecond ordered ultrashort pulses of infrared radiation. Photons that are transmitted in vivo and remain relatively close to ballistic arrive at the light receiving sensor first, while quasi-ballistic photons arrive next, and the strongly diffuse photons arrive last. The variation in the arrival time is caused by the diffusion of light when transmitted; by comparing the differences in arrival time predicted based on the light diffusion equation to the actual arrival times, the Hb concentration can be calculated. This method is called time-resolved spectroscopy, or TRS.
（Image: Hamamatsu Photonics K.K.）
In the initial studies was only effective for a limited area. In the study of Washington University, the new device contains 96 sources and 92 detectors and carries a temporal resolution of 10 Hz. Distances between the source and detector range from 1.3-4.7cm, but what makes this device novel compared to previous iterations of the H-D DOT system is that it covers a much larger surface area from the frontal cortex to the temporal cortex all the way to the visual cortex.
As verification, the researchers obtained four measures through H-D DOT and compared the results to those observed through fMRI. These measures included: examining activity in auditory areas when words were heard, examining activity in the visual areas when reading words, measuring brain activity while imagining speaking and during covert verb generation tasks. In addition, resting state connectivity was measured; in all cases, H-D DOT results were highly congruent to the activity observed through fMRI.
Estimates of the Abilities and Limitations of the Openwater Device
Based on our knowledge of H-D DOT we imagine the Openwater device will consist of a similar system made smaller and portable; likely as a wearable hat. We also propose this new device will be more energy efficient. We hypothesize the measurable brain activity will have a 5-15 mm spatial resolution and approximately a 10 Hz temporal resolution. The most notable limitation to the device is that it is unlikely the brain activity measured will be the activity of the neurons themselves, but rather the changes in blood flow and hemoglobin. Additionally, the measurable area will be limited to about 2.5 cm below the surface of the cortex meaning the device will not be able to record activity in deeper parts of the brain.
What Will the Device That Realizes Telepathy Look Like?
So the question remains, is it be possible to read people’s minds and achieve telepathy by measuring changes in hemoglobin at the surface of the cortex with 5-15 mm spatial resolution and 10 Hz temporal resolution?
Jepsen seems to think so and has already described the research she believes will pave the way for this novel technology. We delved further to decide for ourselves.
New Technology Measures Neural Activity Directly Through Optics
The research Jepsen referred to uses neural cells from Aplysia Californica, more commonly known as sea slugs, to determine neuronal shape and activity. In these experiments, the sea slug abdominal ganglia were submerged in artificial sea water and hit with a Ti-Sapphire laser which was centered at 800 nm with an approximate 100 nm bandwidth. These parameters correspond to a 3 micrometer axial resolution. The experiment is based on the premise that the diffusion of light differs between active and resting neural cells and this attribute can be exploited to estimate neuron activity levels.
The study done by Graf et al. in 2009 eloquently demonstrates that optical coherence tomography (OCT) can be used to measure the shape of each neural cell with a spatial resolution in the μm range.
By using optical coherence microscopy (OCM), the researchers showed that the increase in membrane potential is correlated with an increase of the level of light diffusion. The sampling frequency is 1 kHz.
By applying this technology, it may be possible to measure neural activity with spatial resolution in the range of micrometers and a 1 kHz temporal resolution.
Using Phase Conjugation Mirror Technology to Reconstruct Images
Other areas of interest for Jepsen include research on holography and phase conjugation.
Assuming that objects transmit light, the light phases usually deviate and diffuse. Remarkably, using a special mirror called a phase conjugation mirror, the original object image can be reconstructed by reflecting the light without deviation.
Better Resolution than DOT?
While Jepsen did not discuss alternatives to DOT in her presentation, it should be noted there is an MIT-driven project called All Photon Imaging which has shown increased spatial and temporal resolution compared to the DOT devices.
The MIT group uses a 15mm thick “phantom” as an imitator of the human body and their goal is to capture images occurring behind the body. The object is hit with a diffused titanium-sapphire laser and light transmitted through the phantom is observed with a camera. Similar to H-D DOT, All Photon Imaging characterizes laser beams based on their diffusivity and the speed at which they are detected. Recordings are taken every 2 picoseconds and once again TRS is used to reconstruct the object image.
In this experiment, when using a slit for the subject, it was possible to resolute a 1 cm spatial resolution, but the system was unable to resolve a distance of 5 mm.
By assessing the structural similarity index (SSim), the researchers were able to compare the reconstructed image to the original; based on the high SSim score of 0.78, the images were shown to be highly correlated. For comparison, images generated based on ballistic photons alone have SSim scores of around 0.17-0.27.
Photo-acoustic and Acousto-optic Technologies
Lastly, single-impulse panoramic photoacoustic computed tomography (SIP-PACT) is an example of non-invasive technology offering good spatial and temporal resolution.
This technology is capable of measuring structures and functional images up to a depth of 48 mm with a spatial resolution of 125 μm, and a temporal resolution of 50 Hz.
What Abilities Would the Final Product Possibly Have?
The technologies described above are currently individual devices; in the future, by combining their abilities it may be possible to accomplish the following:
- Measure actual neural activity (i.e. the change in the action potential or the membrane potential), without using the change of blood flow or hemoglobin as indirect measures
- 5 mm spatial resolution
- 1 kHz temporal resolution
- Wider measurable areas (at least the surface of cortex)
Will Telepathy Actually be Realized?
If we can measure neural activity non-invasively, with high spatial and temporal resolution, will we be able to read people’s minds and achieve telepathy?
Openwater is recruiting software architects to statistically estimate thoughts based on recorded changes in oxygenated and deoxygenated hemoglobin. This is intimate information humans have not had access to in the past, therefore it remains to be seen if mind reading and telepathy are possible. Currently, there are studies using fMRI to estimate activity in the visual areas and ECoG experiments which predict conversations based on activity in the frontal and temporal lobes. Furthermore, both EEG and TMS have been used to communicate “yes” or “no” through the internet.
These preliminary findings through fMRI and EEG may be the humble beginnings on the path towards telepathy. If newer, more sensitive, devices can be invented, it would provide a monumental opportunity to achieve telepathy.
Want to be a part of the pursuit for telepathy?
Currently, there are 4 job postings listed on the Openwater website (as of October 2017). For those interested Openwater is looking for candidates with the following skills:
- TFT Designer/Architect – Small Pixels
- Detector Architect
- Ultrasonic/Acousto-Optic Physicist
- Software Architect
The TFT Designer’s responsibility is to construct the LCD or OLED infrared light. The Detector Architect position entails measuring light and constructing the high-resolution CMOS sensor. These two positions are basic and necessary to minimize the emission and reception of infrared light. While Jepsen did not outline the tasks conducted by the Ultrasonic/Acousto-Optic Physicist, it is likely the position would involve testing technology using light and ultrasound as well.
- Adam T. Eggebrecht, Silvina L. Ferradal, Amy Robichaux-Viehoever, Mahlega S. Hassanpour, Hamid Dehghani, Abraham Z. Snyder, Tamara Hershey & Joseph P. Culver. Mapping distributed brain function and networks with diffuse optical tomography. Nature Photonics 8, 448–454 (2014)
- Benedikt W. Graf, Tyler S. Ralston, Han-Jo Ko, and Stephen A. Boppart. Detecting intrinsic scattering changes correlated to neuron action potentials using optical coherence imaging. Optics Express Vol. 17, Issue 16, pp. 13447-13457 (2009)
- G. Satat, B. Heshmat, D. Raviv and R. Raskar. All Photons Imaging Through Volumetric Scattering. Scientific Reports 6, Article number: 33946 (2016)
- Li, L. et al. Nature Biomedical Engineering 1, 0071 (2017)