Prefrontal cortex

Prefrontal cortex

See also

frontal lobes

Aka

PFC

Part of

working memory

cognitive control

See also

analogy-making review paper stuff from green et al. on BA10, and email from martin monti

Consists of

dorsolateral prefrontal cortex (BA 9, 46)

inferior or ventral prefrontal cortex (BA 11, 12, 13, 14)

orbitofrontal cortex (BA 11, 13, 14)

aka ventrolateral prefrontal cortex???

medial frontal cortex (BA 25, 32)

sometimes considered part of the anterior cingulate region rather than prefrontal cortex

ventromedial prefrontal cortex

cognitive control over emotional responses

strong connections to limbic system (amygdala)

cf Phineas Gage

2 multimodal areas of the frontal lobe = the lateral premotor cortex (area 6) and area 46

Mechanism

prefrontal areas = roughly endpoints of the dorsal + ventral visual streams

Felleman & van Essen included the prefrontal as part of the visual cortex

Miller & Cohen (2001)

PFC is not critical for simple, automatic behaviors

stroop task

frontal impairment -> have difficulty

especially when the instructions vary frequently

Wisconsin card sort task

Aka

WCST

Definition

sort cards according to

shape

color

number of symbols

sorting rule changes every so often

so no single stimulus-response mapping will work

PFC damage

able to acquire initial mapping

unable to adapt when the rule changes (Milner, 1963)

PFC in cognitive control

active maintenance of representations of goals and the means to achieve them

provides bias sigals for the rest of the brain

resolves competition

guides activity along appropriate pathways

establishes mappings needed to perform tasks

monkey PFC

grouped into regional subdivisions

orbital

(ventro???)medial

lateral

sensory

mid-dorsal

sensory

partly unique but overlapping connectivity patterns with the rest of the brain - suggest some regional specialization

sustained activity

Fuster (1971)

PFC neurons remain active after transiently presented cue until execution of delayed response

can be specific to the type of information

location/identity of stimulus

action to be performed

expected rewards

or more complicated properties

etc.

and in the face of task-irrelevant distractors

useful for associative learning when one of the stimuli is no longer present

midbrain dopaminergic neurons

fairly low spontaneous activity

bursts of activity to behaviorally salient events

especially unpredicted, desirable stimuli (Mirenowicz & Schultz, 1994, 1996)

activate progressively earlier in time as learning progresses

by events tha tpredict reward

don't fire for the now-expected reward

if the predicted reward doesn't appear

activity is inhibited at the expected time of its delivery

if the reward comes earlier than expected

also get dopamine responses

seem to be coding prediction error

= the degree to which reward (or a cue associated with reward) is surprising (Montague et al, 1996; Schultz, 1998)

provides information that helps you expect a reward

and therefore helps the PFC learn to guide behavior to achieve it

frontal patients

disturbances in learning and decision-making when the evaluation of reward is involved (Rolls 2000)

PFC is modulatory rather than transmissive

transmission = the pathway from input to output runs through PFC

not the case

modulation = guides activity flow along task-relevant pathways in posterior/subcortical areas

like a switch operator in a system of railroad tracks (linking stimuli to responses)

frontal patients

components of a complex behavior are left intact

unable to coordinate them in a task-appropriate way

e.g. stirring then added the milk when making tea

well-practiced tasks spared, new learning impaired

so

you should often get posterior activity without PFC

i.e. transmission without modulation

but not vice versa as often

should get more PFC activity in 'controlled' processes, e.g. color naming

which should diminish with practice

see controlled vs automatic

working memory

traditional distinction between storage and executive (Baddeley, 1986)

they argue

executive control = active maintenance of a particular type of information: the goals and rules of a task

fits with production system models like ACT*

attention and inhibition = same underlying mechanism - guiding activation (Desimone & Duncan, 1995)

according to the biased competition model

inhibition = local competition, rather than centrally by the PFC

binding function of selective attention (Treisman & Gelade, 1980)

PFC is selectin gthe desired combination of stimulus features to be mapped onto the response over other competiting features

PFC vs hippocampus

PFC

extracts regularities in goals and rules across episodes

activity-based control

hippocampus

specific episodes

weight-based control

i.e. hippocampus lays down new tracks. PFC switches between them

PFC damage

perseveration

inadequate updating

distractibility

inappropriate updating

use dopamine to learn when/what to gate in

can this support hierarchical subgoaling and sequences?

PFC

30% of cortical mass

old ideas about functional organization

could be organized into:

• behavioral inhibition vs memory
• sensory vs motor
• based on stimulus dimensions
• sequential order

or by region:

• orbital and medial

  behavioral inhibition

• ventrolateral and dorsal regions

  memory or additional functions (Fuster, 1989; Goldman-Rakic, 1987)

• ventral

  maintenance of memory

• dorsal

  manipulation of information (Owen et al, 1996)

their ideas about functional organization

based on ideas about the biasing signals provided by different regions

orbital

associated with social, emotional and appetitive stimuli

'hot' stuff

i.e. more reflexive, could give rise to possibly inappropriate behavior

so orbital may be biasing against (i.e. inhibiting) those prepotent behaviors

more-dorsal

cognitive, 'cold'

less likely for some responses to be massively prepotent, so the competition between them will be less fierce

apparently the Rougier et al model breaks down the distinction between inhibitory and memory processes within the PFC

argues for a single processing mechanism operating over different types of representations

active maintenance

cellular

neuron bistability - biophysical properties of individual cells

transitions between states are triggered by inputs to the PFC but maintained via the activation of specific voltage-dependent conductances (Wang, 1999)

circuit-based

recirculation of activity through recurrent loops or attractor networks (Hopfield, 1982) -> self-sustained activity (Zipser, 1993)

could be intrinsic to the PFC (Pucak et al, 1996; Melchitzky et al, 1998)

or might involve loop through other brain structures (Alexander et al, 1996), e.g.

prefrontal cortex

striatum

globus pallidus

thalamus

prefrontal cortex

differentiate between the capacity limits of

cognitive control

inherent limit on the number of representations that can be actively maintained and kept independent of one another within an attractor network (Usher & Cohen, 1999)

short-term memory (Miller, 1956)

sensory inputs

dorsolateral = 8, 9 and 46

ventrolateral = 12 and 45

8, 9, 12, 45

inputs from visual, auditory adn somatosensory cortex

9, 12, 45 and 46

inputs from the rostral superior temporal sulcus

bimodal or trimodal responses

arcuate sulcus region (8 and 45) and 12

particularly multimodal

in all these cases, PFC is connected with secondary/association cortex, rather than primary cortex

motor outputs

dorsal PFC (esp 46)

preferential connections to motor areas

46

sends to

motor areas in the medial frontal lobe

supplementary motor area

pre-supplementary motor area

rostral cingulate

premotor cortex on the lateral frontal lobe

cerebellum and superior colliculus

8 (frontal eye fields)

no direct connections between PFC and primary motor cortex

but there are extensive connections with premotor cortex

which send to primary motor cortex and spinal cord

dense interconnections between PFC and basal ganglia

limbic connections

orbital and medial

closely associated with medial limbic structures

direct and indirect (via medial dorsal thalamus) connections with the hippocampus, amygdal and hypothalamus

intrinsic connections

most PFC regions rae interconnected with most other PFC regions

interconnections between all three major subdivisions

ventromedial

lateral

mid-dorsal

also between their constituent areas

lateral PFC is particularly well-connected

ventrolateral areas 12 and 45 interconnected with

dorsolateral areas 46 and 8

dorsal area 9

ventromedial areas 11 and 13

Prefrontal cortex and memory

from Chris Chatham's PFCandLTM1.ppt

PFC damage (Simons & Spiers, 2003)

deficits to source memory and contextual details

impaired ability to resolve interference between to-be-recalled items

confabulation

but intact recognition

levels of processing framework

deep encoding should engage PFC more (Kapur et al, 1994)

semantic processing engages more anterior regions than phonological processing (Poldrack et al, 1999)

how is this related???

left frontal cortex for maintenance (Gabrielli et al, 1998)

or could be selection demands

subsequent memory. frontal activity at encoding predicts

verbal memory accuracy - left VLPFC (Wagner et al, 1998)

visuospatial memory accuracy - right DLPFC (Brewer et al, 1998)

relative strength of these memories (Henson et al, 1999)

VLFC more active in cue-specified retrieval

DLFC more involved in retrieval monitoring

FC activity does not predict recognition accuracy

summary (Burgess & Shallice, 1996)

DLPFC

monitoring and verification of retrieved information

APFC

higher-level mnemonic control operations

VLPFC

cue-specification, strategic search of MTL stored representations

maintenance of stored information

MTL

comparison of retrieval cue and stored representations using pattern completion

source memory

more cue specification

more recollection monitoring

tip of the tongue phenomenon (Maril, Wagner & Schacter, 2001)

state of conflict between metacognition and cognition

ACC and PFC activity

References

Rougier, Noelle, Braver, Cohen & O'Reilly (2005)

From Oxford notes

Prefrontal cortex

prefrontal lobes = form the largest single division of the cortex in humans

Connections

diverse output:

extends to the hypothalamus as well as to the striatum, subthalamus and midbrain

receives afferents from:

the correspondingly large dorsomedial nucleus of the thalamus

(which receives from the frontal lobe, but also the hypothalamus and other parts of the limbic system)

Phineas Gage

"fitful, irreverent, indulging at times in the grossest profanity (which was not previously his custom), manifesting but little deference for his fellows, impatient of restraint or advice when it conflicts with his desires, at times pertinaciously obstinate, yet capricious and vacillating"

his friends even said that he was actually happier: more carefree + less inhibited afterwards

Experiments on animals

lesions in the frontal lobes seem to > lower anxiety monkeys worry less when they make mistakes in learning tasks thought it might help schizophrenics or depressive patients: 1935 frontal leucotomy

pharmacological agents (more reversible) in 1960s

alleviation of tension + anxiety, better adjustment to work (???), increased weight + energy

sometimes: changes of personality too far (euphoria, tactlessness, lackadaisical approach, lack of social inhibitions)

helped with intractable pain - not analgesia, but loss of the 'affekt' of the pain, its unpleasant/emotional quality

'Oh doctor, it's absolutely appalling, unbearable'

• yet smiling, and apparently not really feeling it despite being able to sense it

 

only minor effect on ordinary intelligence, except:

difficulties in carrying out more than one program of activity simultaneously

inability to organise actions in proper temporal sequence, e.g. trying to prepare a meal

e.g. monkeys, delayed reaction test

monkey behind glass partition in cage

shown a reward in one of two boxes, then both closed

interval of 10 minutes - partition raised

normal monkeys go to the correct box to receive reward

frontal lesion animals: cannot, unless they spend the waiting period concentrating single-mindedly on the correct doors

unit recordings in prefrontal areas during delayed response trials indicate that these are areas are in some sense 'waiting to do something'

activity in many units starts up on receipt of the command, then firing is sustained until the response is finally made

= defects in the ability to store a program of action for deferred use

anxiety = side effect of the sense that something has to be done in the future

lack of anxiety sometimes = lack of forethought

similarly, by stripping pain of its significance and meaning for the future, we also relieve its emotional threat

Dorsal prefrontal cortex (areas 46 and 9)

Summary pg 153

Areas 9 and 46 receive their main input from the parietal lobe which processes information about the animal itself, and about the space in which it moves and manipulates things.

Monkeys with lesions in area 46 fail to learn delayed response tasks. These are conditional tasks on which the animal must choose between locations on the basis of information in working memory. These impairments can be demonstrated on an oculomotor version of the DR task on which monkeys must direct their eye movements on the basis of locations in which they recently saw a spot of light. During the delay on this task, many cells in area 46 change their activity selectively according to the location of the target.

Monkeys with lesions in area 9 and 46 are impaired at selecting between objects on the basis of their past responses, and also at generating a series of actions. In PET scanning expeirments with human subjects, the dorsal prefrontal cortex is activated when the subjects generate a series of actions at will. In patients there is also a relation between psychomotor retardation and a decrease in regional cerebral blood flow the dorsal prefrontal cortex. This suggests a role for the dorsal prefrontal cortex in generating actions.

Ventral prefrontal cortex (areas 11, 12, 13 and 14)

Summary pg 170

The ventral prefrontal cortex receives a multimodal input from the temporal lobe. Monkeys are impaired at learning what response to make, irrespective of the modality of the cue. There is also evidence suggesting that it may not be essential that there is a delay between the presentation of the cue and the opportunity to respond; however, this evidence is not conclusive. It is argued that the ventral prefrontal cortex selects the goal - e.g. an object - given the current context.

When monkeys learn visual concurrent discriminations, they can solve the problems by learning only about the associations between the stimuli and reward. Monkeys with ventral prefrontal lesions can learn such problems at a normal rate. Furthermore, when human subjects make perceptual judgements, there is no activation in the prefrontal cortex.

The ventral prefrontal cortex is heavily interconnected with the amygdala. Monkeys will learn to deliver rewarding stimulation to the orbital cortex or to deliver rewarding drugs. It is argued that the connections between the ventral prefrontal cortex and the amygdala are involved in the process by which responses are selected on the basis of their success.

Basal ganglia

Summary pg 201

The prefrontal cortex can influence the premotor cortex by cortico-cortical connections and projections through the basal ganglia. Monkeys are severely impaired at relearning a visual conditional motor task if lesions are placed in the ventral thalamus so as to disrupt the influence of the basal ganglia on frontal cortex.

There are cells in the basal ganglia that fire well before movements when monkeys are repeating a movement from memory or deciding what movement to make.

The premotor cortex also interact with the cerebellum via the ventral thalamus. Patients with cerebellar pathology are slow to learn conditional tasks. There is also activation of the lateral cerebellar cortex when subjects habitually produce the same words in response to cue words.

Notes - Fuster, 'The Pre-frontal cortex'

Introduction

prefrontal = cortex of the anterior pole of the mammalian brain

Unitary function

whatever the criteria for tracing its boundaries, no demarcation can be said to outline a structural entity with unitary function

on morphological grounds alone: thanatomical complexity (especially in higher animals), makes its functional homogeneity implausible

behavioural study of animals with selective lesions of this cortex > rules out such homogenity

untiary role: also inconsistent with clinical findings in patients with injuries to this part of the brain large number of diverse + seemingly unrelated facts - apparently multiople functions - but het basic funcitons seem to be essentially few, and are represented over the cortical surface according to a certain topological pattern interrelated, mutually supporting and complementing functions in the purposive behaviour of the organism prefrontal - ugly, misuses 'pre', aka frontal granular cortex (cytoarchitectonic features in primates) and frontal association cortex (ambiguities of the word 'association') often referred to as 'frontal', implicitly excluding the motor and premotor cortex in rodents and carnivores, is also called the 'orbitofrontal cortex', easily confused with 'orbital frontal cortex' (which in primates the ventral aspect

of the frontal lobe which forms part of the prefrontal cortex)

defined as the part of the cerebral cortex that receives projectisons from the mediodorsal nucleus of the thalamus (applicable to all mammalian brains)

unitary function - but at different levels

Chapter 8 - Overview of prefrontal functions

Summary

primates: cerebral cortex of both hemispheres is divided by the central sulcus (Rolandic fissure) into 2:

posterior - sensation, perception, perceptual memory

frontal - action and motor memory

both are hierarchically organised in terms of development, connectivity, memory and processing of sensory and motor information

dorsal and lateral frontal cortex - segregated action domains for:

• skeletal movement
• eye movement
• speech

  actions are represented by increasing order of complexity + novelty in higher interconnected areas

  abstract schemas = gestalts of actions + goals; novel plans, structures of behaviour

  automatic + routine actions are represented in lower levels of motor hierarchies

  plans: motor hierarchy in the dorsolateral frontal cortex:

  connectivity flows downwards from prefrontal -> premotor -> premotor

  all stages within each action domain are reciprocally connected, as well as with each other through subcortical loops through the basal ganglia

  sequential action: parallel + serial processing

orbitomedial frontal cortex - action domain for emotional behaviour + visceral manifestations

transmits information of limbic origin about the internal milieu -> dorsal cortex

plays a role in decision-making

important cortical depository of emotional memory

frontal lobe cortex - initation and execution of deliberate actions

'executive' functions - decision-making, attention, planning and working memory

= phenomena of neural processing, without unique locations of their own

organism's basic drive + motivations

arrive in frontal cortex from diencephalic and limbic formations

other inputs from sensory receptors and areas of the posterior cortex

attention = ability to select sensory inputs and actions, and to inhibit others

widely distributed in the frontal cortex

dorsolateral = selective

orbital = exclusionary/inhibitory

perception-action cycle = circular flow of organism-environment interactions

sensory processing + consequent action

in cognitive + emotional behaviour

highest level: cycle completed by reciprocal connections between posterior association and prefrontal cortex

prefrontal - mediates cross-temporal contingencies

i.e. bridges time gaps in a structure of behaviour

3 temporal integrative functions of the prefrontal cortex:

1. working memory / active short-term memory

   = the provisional retention of (sensory or motor) information for prospective action

   mainly a function of the action domains of the dosolateral prefrontal cortex

   maintained active in neuronal networks by reverberation through reentrant circuits

2. set

   i.e. motor attention = selection of particular motor acts (from an established repertoire of motor memory) and preparing the sensory/motor systems for them

   essential for execution of plans (temporally extended set)

   also based in the dorsolateral prefrontal corte - though probably under influences from the anterior medial cortex

3. inhibitory control

   exclusionary role of attention

   i.e. protects behavioural structures from external/internal interference (e.g. similar but inappropriate sensory/motor memories)

   based primarily in the orbitmedial prefronal cortex - exerted on a variety of cortical + subcortical regions

Other models of prefrontal function

Cognitive models

Network models

Emotional behaviour

Notes - Rolls, 'Brain & Emotion'

Chapter 4 - The neural bases of emotion

pg 129

prefrontal lobotomies, pioneered by Moniz (Moniz, 1936; Fulton, 1951) - argued that anxiety, irrational fears and emotional hyperexcitabilty in humans might be treated by damage to the frontal lobes

widespread use of this procedure - although irrational anxiety or emotional outbursts were sometimes controlled - but intellectual deficits and other side effects were often apparent (Rylander, 1948; Valenstein, 1974)

still had pain, but it no longer bothered them (Freeman & Watts, 1950; Melzack & Wall, 1996)

Pre-frontal

Notes - Neuroimaging branching study in Nature

Using imaging technology, scientists from the National Institute of Neurological Disorders and Stroke (NINDS) found that a specific type of multitasking behavior, called branching, can be mapped to a certain region of the brain that is especially well developed in humans compared to other primates. The study will appear in the May 13, 1999, issue of the journal Nature.1

"The results of this study suggest that the anterior prefrontal cortex, the area of the brain that is most developed in humans, mediates the ability to depart temporarily from a main task in order to explore alternative tasks before returning to the main task at the departed point," says Jordan Grafman, Ph.D., Chief of the Cognitive Neuroscience Section at the NINDS and a co-author of the study.

The investigators used functional magnetic resonance imaging (fMRI), which measures changes in blood flow to the brain, to view the brains of volunteers while they performed branching tasks. The region of the brain that is involved in multitasking is called the fronto-polar prefrontal cortex (FPPC).

Tasks performed by the volunteers involved exercises to test working memory, attentional focus, and a combination of the two. All of the subjects, who were healthy, normal volunteers, participated in all of the task groups. The task groups consisted of a control task, a delayed-response task, a dual-task, and a branching conditions task. Dual-task involves changing focus between alternative goals successively. The investigators predicted that subject performance on the individual delayed-response task and dual-task conditions would not activate the FPPC. They did predict that the branching task which involves problem solving and planning would stimulate activity in the FPPC. According to the fMRI data, their predictions were correct. The FPPC was activated only during those tasks that involved an interaction between working memory and attentional focus decisions.

The FPPC is the region of the brain that controls complex problem solving and is especially well developed in humans as compared to other primates. The study showed that the FPPC selectively mediates the human ability to multi-task.

Abstracts

Braver & Barch - Common and selective prefrontal cortex regions engaged by working memory and intentional encoding

Functional magnetic resonance imaging (fMRI) was used to examine the role of the prefrontal cortex (PFC) in both long-term memory (LTM) encoding and working memory (WM) tasks involving a variety of material types (words, faces, and pictures). Encoding was studied in a task requiring intentional memorization of items for a later recognition test. WM was studied in the two-back condition of the n-back task. Bilateral PFC in the inferior frontal gyrus (IFG) was found to be jointly activated in both encoding and WM. This region also showed material-specific lateralization in both tasks, with the left hemisphere more active for words and the right hemisphere more active for faces. PFC regions were also found that were selective to either encoding or WM. Right dorsolateral PFC was selectively activated during WM, but showed no material-specificity, while left anterior PFC was selectively activated during encoding of faces and pictures. Activity in medial temporal lobe was also observed, with the left hemisphere engaged by both memory tasks, and the right hemisphere showing significant activity only during encoding. The finding of PFC regions jointly activated during both encoding and working memory tasks suggests that these regions may subserve cognitive processes important for both short-term active maintenance and long-term memorization. Conversely, the finding of selective activation in specific PFC regions and medial temporal lobe suggests that these brain areas may be functionally specialized. Moreover, the results indicate that a complete characterization of the cognitive functions performed by PFC and other brain areas will be best served by integrating findings across multiple memory domains.

Braver & Barch - Mechanisms of Cognitive Control: Active Memory, Inhibition, and the Prefrontal Cortex

Previous research has identified the prefrontal cortex (PFC) as a brain region that is critical for cognitive control. Currently, theorists remain divided about whether to view the PFC as primarily a coordinative, mnemonic, or inhibitory structure. A theory is presented that attempts to resolve some of the apparent conflicts between the predominant views on PFC control functions. In this theory, PFC is proposed to actively maintain representations of context information. These maintained representations provide a mechanism of control by serving as a top-down bias on the local competitive interactions that occur during processing. As such, it is suggested that PFC performs both mnemonic and inhibitory functions in the service of control, and that each is preferentially observable under different task situations. A series of behavioral, computational, and neuroimaging studies are presented that demonstrates how this theory can account for a wide range of data associated with performance of a simple cognitive control paradigm.

Preuss, T. M., 1995.  Do rats have prefrontal cortex?  The Rose-Woolsey-Akert program reconsidered.  Journal of Cognitive Neuroscience,7:1-24.

Primates are unique among mammals in possessing a region of dorsolateral prefrontal cortex with a well-developed internal granular layer. This region is commonly associated with higher cognitive functions. Despite the histological distinctiveness of primate dorsolateral prefrontal cortex, the work of Rose, Woolsey, and Akert produced a broad consensus among neuroscientists that homologues of primate granular frontal cortex exist in nonprimates, and can be recognized by their dense innervation from the mediodorsal thalamic nucleus (MD). Additional characteristics have come to be identified with dorsolateral prefrontal cortex, including rich dopaminergic innervation and involvement in spatial delayed-reaction tasks. However, recent studies reveal that these characteristics are not distinctive of the dorsolateral prefrontal region in primates: MD and dopaminergic projections are widespread in the frontal lobe, and medial and orbital frontal areas may play a role in delay tasks. A re-evaluation of rat frontal cortex suggests that the medial frontal cortex, usually considered to homologous to the dorsolateral prefrontal cortex of primates, actually consists of cortex homologous to primate premotor and anterior cingulate cortex. The lateral MD-projection cortex of rats resembles portions of primate orbital cortex. If prefrontal cortex is construed broadly enough to include orbital and cingulate cortex, rats can be said to have prefrontal cortex. However, they evidently lack homologues of the dorsolateral prefrontal areas of primates. This assessment suggests that rats probably do not provide useful models of human dorsolateral frontal lobe function and dysfunction, although they might prove valuable for understanding other regions of frontal cortex.

Oxford notes

Functions of the prefrontal cortex

prefrontal cortex: controls the cognitive processes so that appropriate movements are selected at the correct time + place

this selection may be controlled by internalised information, or may be made in response to context

the internalised record of what has just occurred is independent of the existing sensory information = the STM

temporal memory = neural record of recent events

events = either things or places

thus information is derived from the object-recognition or spatial streams of sensory processing

(both project to the prefrontal cortex, though to different parts)

i.e. spatial + object information are stored in temporal memory - but localised in different places in the frontal cortex

dorsolateral areas = especially involved in the selection of behaviour based on temporal memory (if defective, become dependent on environmental cues)

so frontal lobe injury => difficulty inhibiting behaviour directed to external stimuli, as opposed to being controlled by internalised knowledge

---