FERRO, Marc, Soft Brain Computer Interface BCI, Melosh Lab, Feb 28, 2019
<p>FERRO, Marc, Soft Brain Computer Interface BCI, Melosh Lab, Feb 28, 2019</p>
<p> </p>
<p>this is on Thursday, Feb 28, 2019 at Stanford in Applied Physics.</p>
<p> </p>
<p>BMI interface (brain machine)</p>
<p> </p>
<p>challenges: Pia is tough, flexible; need > 1 mm.</p>
<p> </p>
<p>current brain implants:</p>
<p> </p>
<p>cochlear: only 36 trodes;</p>
<p> </p>
<p>medtronics: 8 trodes;</p>
<p> </p>
<p>Neuropace: 4 trodes to detect a seizure;</p>
<p> </p>
<p>Galvani device...</p>
<p> </p>
<p>see Seymour review in Microsystems and Nanoengineering in 2017 3, 16066.</p>
<p> </p>
<p>Utah Array;:</p>
<p> </p>
<p>the only pre-clinical device:</p>
<p> </p>
<p>100 trodes; each has a platinum tip and parylene shaft.</p>
<p> </p>
<p>it uses a connector pedestal with 100 pins and uses a wire bundle.</p>
<p> </p>
<p>it requires external attachments; </p>
<p>the signals degrade over time;</p>
<p> </p>
<p>foreign body reaction; get bleeding and reactive tissue;</p>
<p> </p>
<p>inflammation: from reactive astrocytes and cytokines.</p>
<p> </p>
<p>can destroy existng networks;</p>
<p> </p>
<p>chronically implanted rats exhibited up to 527% increase in time to perform a fine motor task.</p>
<p> </p>
<p>the brain BEATS in time with heart and the respirations. and when you move;</p>
<p> </p>
<p>surface micromotion + 10 to 30 microns due to resps and</p>
<p>2 to 4 microns due to vascular pulsatility.</p>
<p> </p>
<p>Rogers and Litt groups</p>
<p> </p>
<p>flexible silicon CMOS arrays</p>
<p> </p>
<p>OECT = organic electrochemical transistor embedded</p>
<p>in ultrathin organic film</p>
<p>as in Malliara and Kohagholi groups</p>
<p> </p>
<p>NeuroPixels as in Harris group at IMEC.</p>
<p> </p>
<p>Neuro-Mesh is a long term stable polymer</p>
<p> </p>
<p>see Salatino et al in 2017 Nat Biomed Eng review.</p>
<p> </p>
<p>Young's modulus: Gigapascal to v small.</p>
<p> </p>
<p>brain tissue is about 10-5 GPascal.</p>
<p> </p>
<p>stress over the strain.</p>
<p> </p>
<p>but that's difft than bending stiffness.</p>
<p> </p>
<p>so, it depends on the geometry of the implant.</p>
<p> </p>
<p>proven with microwire trodes</p>
<p> </p>
<p>that are mostly 7 micron carbon.</p>
<p> </p>
<p>smaller is better.</p>
<p> </p>
<p>what is force required to punture and penetrate the brain...</p>
<p> </p>
<p>initally goes up at the pia but then it gives, and then get</p>
<p>friction force as it penetrates.</p>
<p> </p>
<p>shows difft trodes from 100 micron width and down to 7.5 microns (much better)</p>
<p> </p>
<p>get no visible bleeding with a 25 micron microwire!!! (shows video with stained vasculature.</p>
<p> </p>
<p>considerable reduction of initial stab injury.</p>
<p> </p>
<p>this led us to</p>
<p> </p>
<p><strong>NeuroRoots:</strong></p>
<p> </p>
<p><strong>goals are long-lasting; high channel count;</strong></p>
<p><strong>single cell recordings
reliable day after day</strong></p>
<p> </p>
<p>shows human axons.... abt 10 micron in diamter;</p>
<p> </p>
<p>we made 10 micron microwire tips (a sphere on the tip of a thinner wire.</p>
<p> </p>
<p>made out of platinum.</p>
<p> </p>
<p>our neuroroots are 40X stiffer than an axon.</p>
<p> </p>
<p>but we can match a human modulus of 1.5 micron diameter.</p>
<p> </p>
<p><strong>length (depth of 600 microns)</strong></p>
<p> </p>
<p>sets of tetrodes;</p>
<p> </p>
<p>but hard to implant cuz so thin and flexible.</p>
<p> </p>
<p><strong>use a 35 micron needle and the trodes snap onto the needle.</strong></p>
<p> </p>
<p>he shows the trodes adhereing to the needle.</p>
<p> </p>
<p>capillary force....</p>
<p> </p>
<p>then perhaps add a little PEG as temporary glue;</p>
<p> </p>
<p>he shows an oblique penetration (somewhat tangential to surface of brain)</p>
<p> </p>
<p>we've done chronic implantation in rat</p>
<p> </p>
<p>wih Alex Gonzalez from Giocomo Lab.</p>
<p> </p>
<p>freely moving rat over 7 weeks;</p>
<p> </p>
<p>easy to implant;</p>
<p> </p>
<p>stable over 7 weeks... shows action potential with no diminution over 7 weeks.</p>
<p> </p>
<p>shows place cells in a maze.</p>
<p> </p>
<p>next tried to implant in Cerebellum;</p>
<p> </p>
<p>> half of the neurons in the brain;</p>
<p> </p>
<p>they are 4 to 6 microns;</p>
<p> </p>
<p>compare to calcium imaging:</p>
<p>poor signal to noise ratio;</p>
<p>and slow kinetics;</p>
<p>cannot stimulate;</p>
<p> </p>
<p>we did expts with Sriram Jayabal from Jenn Raymond Lab.</p>
<p> </p>
<p>in vitro....</p>
<p> </p>
<p>these are probably Purkinje cells;</p>
<p> </p>
<p>next....</p>
<p> </p>
<p>clinical subdural BCI micro-implant place into sulcus with</p>
<p>Prof Ada Poon in collab with Jaime Henderson...</p>
<p> </p>
<p>place everything below the skull....</p>
<p> </p>
<p>wireless xmission; low power;</p>
<p> </p>
<p>length can be anything....</p>
<p> </p>
<p> </p>
<p> </p>