Transcript optical - Rackcdn.com

Brain Stimulation for Memory
Mijail “Misha” D. Serruya, M.D., Ph.D.
Outline
•
•
•
•
•
Definitions of memory and stimulation
Macrostimulation
Microstimulation
Optical
Challenges
Memory
• Procedural
• Semantic
• Episodic
– Binding of item and context
– Spatial context & navigation
– Temporal context
– Mental time travel
rTMS
Stimulation
Non-Invasive
tDCS
tACS, tSOS
Macro
Electrical
Platinum-iridium, steel
Epidural/Subdural
Depth/DBS/Grids
Low impedance <1kΩ
1-5 mm diameter
0.5 to 15 mA
Optical
Cortical MEA
Depth microwires
100kΩ to 2MΩ
10 to 100μm contacts
50 to 100 μA
Implanted
Chemical
Microfluidics
Micro
Biologic
Stimulation to Elicit Memory
• A seizure as ‘natural’ stimulation
– Perirhinal ‘familiarity’ separate from context
experienced as déjà vu preceding an MTL seizure
• Penfield:
– Cortical macro-stimulation could induce vague
sensations of familiarity or vivid re-living or
recollection of a memory
• Example: Reported patient who would hear orchestral piece
at same tempo as it was originally heard; stopped when
stimulation stopped and would restart at beginning when
stimulation repeated
Basal Forebrain DBS
• Hamani, Laxton, Lozano and colleagues
• Hypothalamic fornix DBS in single patient in attempt to
treat obesity
– Serendipitously provoked reversible retrievable of distinct
autobiographical episodes
– EEG source localized activation to MTL / HF
• In 1969 Heath reported septal autostimulation in humans
• Patient B7: zap awake out of
narcoleptic stupor; sexual
• Patient B10: zap out of psychotic
rage into tranquil bliss
• +4 decades: pilot trial:
θ
• 6 patients with mild or probable AD
• Mean age 62 years old
• Fornix DBS for one year
Basal Forebrain
Septal nuclei
Ventral pallidum
Anterior
commissure
Amygdaloid
complex
Substantia
inominata
(nucleus basali
of Meynert)
Uncinate
fasciculus
Diagonal band
of Broca
fornix
Fornix DBS
Metabolic
benefit vs
cognitive
benefit?
What is
going on?
SS
‘Blind’
open-loop
130 Hz?
Stay tuned
Entorhinal Cortex DBS
• Suthana, Fried and colleagues
• MTL (EC & hipp) macro-stimulation in patients with
• What is going on?
medically refractory epilepsy
• Enhancing salience of
incoming sensory
information into perf
path?
• Opposite: silencing EC
input to enhance CA
auto-association?
• How account for multiple
comparisons?
• How account for
propensity to wander
aimlessly for prolonged
period in ‘yellow cab’ task
In half trials, focal stimulation (below after-discharge
(adding significant noise)?
threshold) applied
• How would it work in real
• Patients reached landmarks that were learned while
world?
accompanied by EC stim more quickly and efficiently than
• Zap yourself when
locations that were not accompanied by stimulation
you park your car to
• Hippocampal stimulation had no such effect
solidify memory of
where you parked it?
• Patients learned destinations in a
virtual environment
•
Cortical Surface Macrostim
• Kahana, Lega, Burke, Jacobs and colleagues
• Verbal instead of spatial memory
• A signature of enhanced encoding
-SME
Free Recall Task
+SME
ball
tree
SME =
Subsequent
Memory
Effect
flag
time
+
tree
baseline
power
gamma
power
increased
compared to
baseline
Leverage the Signature
Electrical Stimulation
ball
tree
flag
time
+
tree!
Leverage the Signature
Closed-Loop Real-time Feedback
ball
tree
flag
time
+
tree!
Targeted Closed-Loop Microstimulation
• Hampson, Deadwyler, Berger and colleagues
• Macrostimulation non-specifically activates larger volumes of gray matter
and white matter fibers of passage in a possibly diffuse, modulatorymanner
• Microstimulation specifically targets smaller volumes of gray matter,
including hippocampal sub-fields, hence could transmit specific information
• Difference between enhancing an already existing circuit versus imparting
an entirely new circuit
Basic Hippocampal Circuit
Schaffer collaterals
Subiculum
Mossy fiber
pathway
Associational
commisure
Dentate
gyrus
•
•
•
Perforant path
Entorhinal
cortex
Long loop:
•
EC2DGGCCA3CA1subiculumEC5
Short loops:
•
EC3CA1EC5
CA3 project mainly to CA1 and to recurrent CA3 population
•
•
•
•
•
It works!
Enhanced learning when activated
Restored learning if MK801 infused chronically
into hippocampus
No restoration if microstim random
But life is not binary DNMS: real-world?
Spike-Triggered Microstimulation
I love
Vegas…
Can’t…
Stop…
Playing…
Optical Stimulation
• While electrical microstimulation is far more spatially specific than
electrical macrostimulation it cannot match the specificity of
optical stimulation of neurons genetically modified with ChR2
family
• Contextual component of fear memories
– Dentate gyrus discriminates between similar contexts
– Sparse (2%) populations of DG granule cells activated in given context
– Although same population of DG cells activated repeatedly in same
enviornment, different environments activate different DG
subpopulations
Optical Stimulation
• Integrated opto-electronic devices are in
development
• Nurmikko, Deisseroth and colleagues:
Prototype Brain Interface: Percutaneous
Signal OUT
skin
Connector
Connector
cap
Bone
Arachnoid
Dura
I
Ventricular III
Shunt
Cortex
3.0 mm V
500 µm
400 µm
VI
White Matter
J. Donoghue 1/2001
Biologic Interfaces
• Cullen, Chen, Wolfe, Smith and colleagues
Even if microelectrodes could record spatially specific
neurons, and even if optical fibers could activate
spatially specific neurons:
– Bandwidth limits
– Nature already maximized
• Hybrid solution?
• Autologous construct
Bioprosthetic Neural Modules
Biological Hysteresis
• Brown & Sherrington, 1912
• March ball electrode back
and forth across macaque
precentral gyrus
• Record EMG from supinator
longus:
• Identical stimulation parameters
at identical ‘cortical point’ causes
different effects if intervening
stimulation elsewhere occurs!
Biological Hysteresis
Change in Oscillatory
z(Power) with Electrical
Macro-Stimulation
• The effect of stimulation is a function of
preceding, endogenous oscillatory power
0
0.2
1
Pre-Stimulation Oscillatory
z(Power)
Given all this implanted hardware…
Scalp (EEG) Electrode
Subdural Grid (ECoG)
Multi-electrode Array (unit)
Intraosseous / Epidural Array
Depth Electrode (ECoG) or DBS
Microwires
…how do we forge a
Cognitive Assistive Device?
Learn facts, faces,
voices
Couple to
camera/microphon
e AI
Reminders for medications,
tasks
Doctor to adjust
connectivity
between brain
regions
Record and ‘replay’
local and global
brain states and
temporospatial
activation patterns
Challenges
• We understand a lot less about human memory than, for example,
Parkinson’s disease / dystonia or epilepsy
• Memory deficits from Alzheimer’s disease, traumatic brain injury and
other etiologies are rarely as ‘lesion-specific’ as the focal nigrostriatal
degeneration in PD
• Open-loop: how physiologic?
– Ignore fine timing: altering precise temporal delay between septal & DG
input to CA3 can mean difference between LTP and LTD / nic v musc
– How different than simply infusing Ach into basal forebrain? Intrathecal
donepezil?
• Modulating an existing network or injecting information?
• How should the loop be closed?
– The physician derives parameters?
– The patient improvises in daily life?
– The device auto-associates?
• Optics: gene therapy in humans?
– Optics for ‘DBS’ mismatch…while optics for memory ideal!
• How take into account biological hysteresis?
• This is an airplane we will build while we are flying it!
FINIS
• STAY TUNED