1998 Milner Goodale Visual Brain in Action

Jan 25, 2014 03:45 PM

this is a precis of the book Visual Brain in Action

standard account:  visual constructs an internal model of the world: a simulacrum of the real thing.

(not disputed)...

even tho this is IMPORTANT  (percep repr of objects);

vision ev olved  not for percpetion bu to provide  distal sensory control of our many movements.

many visual control systems for different motor output evolved as independent IO modules..

eg catching prey, to avoiding obstacles...

human behavior may be arbitrary and clearly mediated by some sort of internal model of world.

systems evolved to permit brain to model the world, to ID objects and evnts etc.  (meaning and CRs)

such systems are not linked to specific motor output but instead to cog systems subserving

memory, semantics, planing and communic.

But ultimate function must be production of adaptive behavior.

complex system: to SELECT appropr courses of action 

versus  immediate visual control requried to execute those actions.

vision for action vs vision for perception...

Section 2:  VISUAL BRAIN:

visual areas occupy posterior 50% of CCortex  (Zeki 93);

ventral stream goes to IT eventually;

dorsal stream to PP cortex = poster parietal

(ungerleider and Mishkin 82.... note Bob: 30 years ago!!!)

but IT and PP also get inputs from Super Collic  (via thalamus)

(see direct connection from retina to SC, then to PUlvinar then to PPCortex (separ from dorsal stream);

1982:  ventral vs dorsal IS 30 years old!!!

object vision  vs spatial vision...  WHAT versus WHERE;

IT lesions:  could not discrim objects but did not affect Landmark task.

More recent accounts:  it IS GUIDANCE of ACTION that PP cortex (dorsal path) does.

Neurons in IT are tuned to features of ojbects  and whether assoc with reward. =>

enduring charactgeristics and signif. of ojects.

PP coretx neurons:  responsive depends on... target of an arm reach;

object of a grasp response;  others:  target of eye saccade;

others when stim is moving, followed by slow pursuit eye mvmt, others whe object is targetof ocular fixation.

and some selective for OEINTATION of an object

important;  many PP neurons respond better when object is ATTENDED in readiness to make a saccade

or manual response.

PP lesions impair landmark task:  monkey cannot orient to landmark.

monkey cannot grasp a moving or stationary food morsel and fail to shape and orient

their finger and hands..

Gallese 97:  inject muscimol into PP cortex:  impaired hand shaping when monkey reaches for grasp obects.

and anatom linked areas of premotor cortex.  Rizzolatti 88.

dorsal stream: primary role in mediating visual control and guidance of acts.

And EGOCENTRIC locations of targets are computed within PP cortex...

done separately for guiding eye mvmts and for mvmts of hands.

Kluver and Bucy 1938:  visual recog knocked out by bitemproal lesions in monkeys.

(but di NOT bump into objects and could still jump)

recent study... monkey could still catch gnat (despite bitemp lesions).

and can track adn seize moving peanut.

same dissoc pub by Balint:  optic ataxia:  difficult reaching for objects;

problem with visuomotor.

lesions in superior parietal lobule and IPS...  diffic executing saccades in space.

also  pts have diffc oritneing hand and forming grasp.

Problems in orienting hand toward a slot....

normal S anticipates size of object grasped.;

Aslo visual FORM agnosia (for IT lesions)

3)  VISUAL AWARENESS:

we suggest... dorsal stream acts ALONE and independent of any acquired knowledge base;

look at visual illusions eg Gregory 97...visual system fooled cuze of stored knowledge.

But dorsal system is NOT deceived by spatial illusions.

dorsal system directs our eye saccades and hands to where the target really is.

We propose that VETNRAL stream generates and is informed by abstract knowledge of objects.

And that coding also rednders the repres. accessible to awareness.

Fits idea that coded descriptions of enduring object properties  (rather than egocentric view)

are what are needed for mental manipulations.

They propose a conjoint requirement for visual awareness:

a) special coidng  (object-based and abstaced away from egocentric particulars  AND

b) above threshold activation (attention).

4)  Visual Brain in Action:

we have emphasized separtion of dorsal and ventral... but multiple connex btwn them

and adaptive goal-directed behavior depends on intergation.

Thus,  execution of goal-directed action depends on dedicated control systems in the

dorsal stem but  SELECTION of approp goal objects and actions in ventral strem.

how do the 2 streams inteeract to create purposve behvior?

visuomotor modules in primate parietal lobe function indep from occipitotemp mechs generating

perception-based know of world.

Dorsal stream is only to guide action for the HERE and NOW... so products are useless for later reference.

Only visual stream can exrcise insight, higdsight and foresight abt visual world.

1998 Milner Goodale Visual Brain in Action

Attachments