Development of a Visual Prosthesis: The Orion Visual Prosthesis System Nader Pouratian, M.D., Ph.D. Daniel Yoshor, M.D., Ph.D. Study partly funded by NIH BRAIN Initiative grant: UH3NS103442 Ex 99.2

Disclosures Second Sight – Grant Support for this project Consultant BrainLab – Grant Support Medtronic – Fellowship Support Boston Scientific – Consultant, DSMB

Overall Goal: Cortical Prosthesis for Previously Sighted Patients with No or Bare Light Perception Safe – Implant-related concerns (infection) Seizures Practical –Surgically Practical and Adoptable (microarray “tiles” vs ECoG) Clinically Useful –Perceptions Shadows Edges

Visual Cortical Prosthesis: Orion I Implant Leverages existing Argus II technology Bypasses the eyes and optic nerve and directly stimulates the visual cortex Early Feasibility Study underway in U.S. Designated a Breakthrough Device by FDA In planning stages for next phase (pivotal or feasibility study)

Orion I Implant Medial View Lateral View Electrode array (60 channels) Cable Implant coil (receiving antenna) Cover Video Processing Unit Glasses & Antenna

Early Testing in a Blind Subject with “Off the shelf” Device 30 year old with an 8 year history of bare light perception blindness due to Voght-Koaynagi-Harada Syndrome Implantation of a Neuropace responsive neurostimulation device with 2 parallel 4-contact leads implanted over the right medial occipital lobe via a posterior interhemispheric approach. Systematic manipulation of stimulation intensity, pulse width, frequency, and site of stimulation over 10 months.

Results: Retinotopic Localization and Reproducibility Over Time Frequency Modulation of thresholds Different shapes and sizes Consistent Localization

Current Thresholds: Relationship to Stimulus Parameters

Differential phosphene modulation with modulation of distinct simulation parameters

Stimulation thresholds are stable over 19 months

Orion Early Feasibility Study: 6 subjects Eligibility: Subjects are completely bilaterally blind due to any cause other than damage to the visual cortex May include trauma to eyes or optic nerve, diabetic retinopathy, glaucoma, or other etiology Primary outcome: safety (adverse events) Secondary outcome: ability to produce phosphenes, long-term functionality of the device, and benefit in terms of visual function, functional vision, quality of life Two centers: UCLA and Baylor College of Medicine

Site Subject Implant Date Age at time of implant Gender Reason for Vision Loss Implant Duration as of 6/28/19 (months) UCLA S1 01/30/2018 56 Male Optic neuropathy secondary to burn/trauma 17 UCLA S2 03/29/2018 29 Male Head Trauma – motor vehicle accident 15 UCLA S3 04/26/2018 64 Female Endophthalmitis secondary to liver abscess 14 Baylor S4 04/27/2018 34 Male Congenital glaucoma 14 UCLA S5 05/31/2018 52 Male Congenital glaucoma 13 Baylor S6 01/17/2019 57 Male Head Trauma–gunshot wound 5 Mean ± Std Dev 48.7 ± 13.9 Ratio F:M 1 : 5 Mean ± Std Dev 13.0 ± 4.0 Orion Early Feasibility Study – Enrollment

6 adverse device or procedure-related adverse events (ADEs) in 2 (out of 6) subjects 1 serious ADE 5 non-serious ADEs Four out of 6 subjects have not experienced any device- or procedure-related Aes Data as of last independent medical safety monitor meeting of May 3, 2019 Primary Endpoint - Safety Event # Subjects # Events Serious Adverse Events 1 1 Seizure 1 1 Non-serious Adverse Event 2 5 Hand Twitch-Bilateral 1 1 Headache 2 2 Visual Aura 1 1 Visual phenomenon 1 1 Grand Total 2 6

Orion System Programming and Home Use After implantation, subjects undergo “programming” process to customize settings for use of the system in video mode (with stimulation driven by the camera image) No earlier than 3 months, subjects are cleared to use the system outside the clinic Visual rehabilitation begins immediately thereafter In the study, a suite of assessments to measure performance is administered at multiple time points Visual function – objective, controlled, artificial Functional vision – subjective, real-world, more meaningful Well-being & quality of life – patient-reported

First Threshold Measurement Subject Average Threshold (µA) Standard Deviation (µA) No. of Thresholds S1 2893 1322 56/60 S2 1645 828 60/60 S3 2194 789 59/59 S4* 3765 1603 54/58 S5 2693 1248 59/59 S6 3730 1170 55/59 Thresholds amount of current required for subject to see phosphene 50% of the time on a single electrode Thresholds have been mostly stable over time * Measured with a modified staircase method

Measuring Ultra-Low Vision Visual Function 20/20 20/200 Bare LP No LP Count Fingers Hand Motion LP with Projection LP = Light Perception Legally Blind Square Localization Direction of Motion Grating Visual Acuity (1.6 – 2.9 LogMAR) (20/800 – 20/16000)

Square Localization Performance * Mean error significantly different ON vs. OFF, t-test, p<0.05 System OFF System ON Subject S2 at 6 months

Direction of Motion Performance * Mean error significantly different ON vs. OFF, t-test, p<0.05 System OFF System ON Subject S2 at 6 months

Grating Visual Acuity is measured between 2.9 – 1.6 logMAR. Gratings are presented for only 5 seconds No subjects (0 of 5) scored on the scale with the System ON or OFF at 6 months post-implant 1 subject (1 of 4) scored on the scale with the System ON or OFF at 12 months post-implant (2.8 logMAR) Training on this task has improved performance considerably Grating Visual Acuity - Performance * Available data are reported as of Jun 2019

Trained observers rate the impact of the Orion implant on patients’ well-being and functional vision periodically at specific time points Interview Observer-rated functional vision tasks (System ON and OFF) Final rating of benefit of Orion System on subject’s life Functional Low-Vision Observer-Rated Assessment (FLORA) Available data are reported as of Jun 2019 Time Point Positive Mild Positive Neutral Mild Negative Negative M6 40% (2/5) 60% (3/5) 0 0 0 M12 75% (3/4) 25% (1/4) 0 0 0

“Real World Use”: observations from rehabilitation sessions Subjects are finding success with Orion for everyday visual tasks “…He was able to find the cue ball with no problems on the table. He was able to tell the cue ball from the blue ten, and also from balls with a stripe vs. the cue ball. He could find the racked balls at the other end of the table too.” “…we've found that looking forward is where [the Orion] shines better for the user as they can detect upcoming objects... He was able to see cars parked on the side of the street, openings in the sidewalk up into driveways, etc.” “…He was able to order patterns from small checkers, big checkers, and white cloth. There's a half inch difference in checker size in the patterns.” “When working throughout her apartment building, she was able to tell where I was located when standing in front of a 10’ wide light wall w/out visual clutter. She was also able to correctly determine whether I was traveling from left to right or right to left along this wall, 7/10 times.”

How does the Orion Visual Cortical Prosthesis System compare to the Argus II Retinal Prosthesis System? Argus II System: retinal implant for end-stage Retinitis Pigmentosa Orion System: potentially treats any cause of profound bilateral blindness other than cortical damage Eye injury, optic nerve disease, glaucoma, retinopathies Excludes reversible conditions such as cataracts Different array size, electrode spacing, and neural placement (diseased retina vs. healthy visual cortex) Electronics case Receiver Antenna Array of 60 electrodes Optic Disk Fovea Tack

Argus II System Indications for Use (U.S.): Bilaterally blind due to retinitis pigmentosa with bare light or no light perception Implanted in the worse-seeing eye Commercial Approval in European Economic Area (CE Mark) in 2011 Commercial Approval in U.S. (HDE) in 2013 Over 350 implanted worldwide (EU, US, Canada, Middle East, Asia) Only visual prosthesis with market approval in U.S. Electronics case Receiver Antenna Array of 60 electrodes Optic Disk Fovea Tack

Safety Comparison Rates of Adverse Events due to the Device or Procedure # Subjects # Events Serious Adverse Events 1 1 Non-serious Adverse Event 2 5 Grand Total 2 6 # Subjects # Events Serious Adverse Events 2 2 Non-serious Adverse Event 4 23 Grand Total 4 25 5 Orion Early Feasibility Study Subjects through 1 Year First 5 Argus II Feasibility Subjects through 1 Year

Performance Comparison Argus II at 12 months Orion at 12 months* Square Localization 94% (N = 16) 80% (4 of 5) Direction of Motion 63% (N = 16) 80% (4 of 5) Grating Visual Acuity 48% (N = 29) 20% (1 of 5) FLORA 80% (N = 15) 100% (5 of 5) Improvement = significantly better ON than OFF (SL & DOM), on scale @2.9 logMAR or better (GVA), or positive/mild positive (FLORA) *6 month data used for 1 or 2 subjects (12 months not measured yet) Percent showing improvement with System ON vs. OFF

Conclusions Visual Cortical Prostheses are promising. Visual cortical prostheses are generally safe. Visual cortical stimulation in previously sighted blind subjects results in consistent and reliable thresholds and phosphenes Surgery is only the first step. The implant will evolve – Need for penetrating? Greater coverage? Learning to use the device are key. Let’s not overlook – this is the highest channel count brain stimulation device with multiple independent current sources