How To Read Brain Images - PowerPoint Presentation

Slide1

How To Read Brain Images

Lasers, Light and Legacy:

Berkeley, Aug 1 2015 Slide2

Into Neuroscience: With

Townes’ blessing

Auditory system (UCSF, 1986-1989)Visual system (HMS, 1989-1992; Rockefeller, 1992-2002; starting faculty job, Columbia 2002)Understanding brain imaging (now: Columbia)

CHT: ‘If I were starting afresh now, I would also study Neuroscience’. Convinced me to start and finish a short PhD project with him. 6 month post-doc while I looked my first Neuro position.

Full circle: CHT coming to my talk at Berkeley 2012Slide3

Functional Brain Imaging (fMRI)

Not directly measuring neural activity

Rather, changes in blood flow, volume, oxygenation etc as a proxy for local neural activity.Assumption: any increases in local blood flow must be due to increased local neural activity. Slide4

Angelo

Mosso

: first ‘brain imaging’ 1880’s

Brain: greedy. 2% body mass but 20% energy.Increased brain activity: pumps fresh blood into the brain, switches it

to where it is needed.Slide5

Angelo

Mosso

: first ‘brain imaging’ 1880’s

Brain: greedy 2% body mass but 20% energy.Increased brain activity: pumps fresh blood into the brain, switches it to where it is needed.Slide6

Basis of BOLD fMRI.

Increased neural activity -> local increase of blood in the brain. Well

establishedBOLD fMRI: turns this logic on its head: M

easures blood flow in the brain‘Where you see increased blood flow there MUST HAVE BEEN local neural activity’.How valid??.Slide7

Basis of BOLD fMRI.

My work shows: at least 2 distinct components to the fMRI response.

1: Stimulus-evoked: driven by local neural activity.2: Task-related / anticipatory:

equally large, but NOT driven by local neural activityBlood flow = local neural activity: not so simple! More nuanced and interesting.Slide8

Where I came in:

Primary visual cortex (V1)

Alert monkeys performing simple visual tasks (fixate, then I flash on a visual stimulus). Intrinsic-signal optical imaging

(optical, high-res version of fMRI) + electrode recording.Yevgeniy SirotinRockefeller U

Karl

Korinek

MPH

Sirotin

& Das

Nature

. 2009Slide9

Surprise!

Sirotin

& Das Nature. 2009Slide10

Sirotin

& Das

Nature. 2009

Dark room (monitor taped over). Fixation task:

Pinhole fixation pt (1-2’ ~ 1-2 cone dia), on all the time, changing color, cue to fixate / relax

(single twinkling star in pitch black night sky)Slide11

Timing Anticipates Trial Onsets

Sirotin & Das

Nature

. 2009Slide12

Timing Anticipates Trial Onsets

Sirotin & Das

Nature

. 2009Slide13

No local neural activity (unlike for visually-evoked response)

Sirotin

& Das Nature. 2009Slide14

V1 Arterial Contraction-Dilation Cycle

Distinct in vein, artery and ‘parenchyma’Slide15

Implications

Implications for brain imaging.

Not all brain imaging signals are born the same.Two neurally distinct imaging signals. Need to be separated.

BUT makes fMRI cleaner: after separation, stimulus-related part 90% predictable from neural activity.Mariana CardosoBruss Lima (UFRJ)Masha BezlepkinaCardoso et al. Nature Neuro

. 2012Slide16

Implications

And the task-related part?

Reward / motivation

Mariana CardosoBruss Lima (UFRJ)Masha BezlepkinaSlide17

Implications

And the task-related part?

Reward / motivationPerformance.

Mariana CardosoBruss Lima (UFRJ)Masha BezlepkinaSlide18

Implications

And the task-related part?

Reward / motivationPerformance.Temporal anticipation

Attention?Mariana CardosoBruss Lima (UFRJ)Masha BezlepkinaSlide19

Where we are now…

Brain imaging: much richer than just a proxy for local neural activity

Design tasks to get at different componentsMechanism: ??? (neuromodulators?)

Grateful to CHT for setting me on this path …