learning & memory in the head direction cell...

Learning & Memory in the Head Direction

Cell Circuit

Jeffrey Taube

Psychological & Brain Sciences

Dartmouth College

Hanover, NH

• Circuitry issues – How the Head Direction signal is

generated?

• Involvement of vestibular & motor cues

• How visual landmark information is processed

• How HD signals guide behavior

• How head direction cells respond in 3-D, as well as under

micro-gravity conditions?

What our Lab studies:

• Circuitry issues – How the HD signal is generated?

• Involvement of vestibular & motor cues

• How visual landmark information is processed

• How HD signals guide behavior

• How head direction cells respond in 3-D, as well as under

micro-gravity conditions?

• Grid cell generation

Review a number of HD cell studies that we have conducted that involve

learning and memory and inform us about some of the underlying neural

processes involved in navigation.

What our Lab studies:

Contributors

Stephane Valerio

Brett Gibson

Ed GolobJeremy Goodridge

Gary Muir Jeffrey Calton

Ben Clark

Others:

Josh Bassett Joel Brown

Paul Dudchenko Russ Frohardt

Jennifer Rilling Mike Shinder

Bob Stackman Sarah Wang

Shawn Winter Ryan Yoder

Two Components of Spatial Orientation

• Location - Place cells - 1971 - John O’Keefe

• Directional heading

But place cells do not provide information

about directional heading.

A second category of spatial cells encodes

for directional heading :

Head Direction cells

- Postsubiculum (dorsal presubiculum)

- found in many limbic system structures

1984 - James Ranck

Head Direction Cell Video

Head Direction Cell Properties

• Direction of head, not body position.

• Head direction in the horizontal plane.

• Fires whether animal is moving or still.

• Firing is independent of location and behavior.

• Each cell exhibits one preferred firing direction.

• Preferred firing directions distributed equally around 360º.

0

°

90

°

180

°

270

°

Peak firing rate

Preferred

firing

direction

Background firing rate

Directional

firing

range

Head Direction (°)

Firing

Rate

(spikes/s)

Where is the Postsubiculum?

M Witter (1989) In Hippocampus: New Vistas

Postsubiculum

- Deep layers

Dorsal portion of Presubiculum = Postsubiculum

Subicular Complex

• Subiculum (S)

• Presubiculum (PRE)

• Parasubiculum (PAR)

3 Typical HD cells from different brain areas

Taube, Muller, Ranck (1990) J Neurosci; Taube (1995) J Neurosci; Stackman & Taube (1998) J Neurosci

Head Direction (°)

# of Samples: 1267

Mean ISI: 12.14

Std. Dev.: 5.07

: 0.418

Info. Content:

2.021

Vector Length:

0.913

ATI:

23.7

Gamma function r:

0.977

Firin

g R

ate

(spik

es/s

)

Head Direction (°)

Time (msec)

Num

b

er

0

50

100

150

0-500-1000 500 1000

C

Interspike Interval (msec)

0 10 20 30 40 50 60

Fre

qu

en

cy (%

)

8

6

4

2

0

10

12

Firin

g R

ate

(spik

es/s

)

# of Samples: 815

Mean ISI: 14.57

Std. Dev.: 10.08

: 0.692

Info. Content: 1.853

Vector Length: 0.838

ATI: 39.8

Gamma function r: 0.933

Num

b

er

0-500-1000 500 10000

25

50

75

8

6

4

2

00 10 20 30 40 50 60 70

Fre

qu

en

cy (%

)B

# of Samples: 635

Mean ISI: 25.01

Std. Dev.: 20.67

: 0.826

Info. Content: 1.481

Vector Length: 0.872

ATI: 75.4

Gamma function r: 0.898

Firin

g R

ate

(spik

es/s

)N

um

b

er

0-500-1000 500 10000

10

20

30

5

4

3

2

1

0

Fre

qu

en

cy (%

)

0 12020 40 60 10080

A

No evidence for

theta-modulation

Inhibitory

Excitatory

Place Cells

Angular Head Velocity Cells

Grid Cells

HD Cells

Head Direction Cell Circuitry

Interpeduncular

Nucleus

?

Nucleus

Prepositus

Medial

Vestibular

Nucleus

Supragenual

Nucleus

Retrosplenial

Cortex

Hippocampus

Postrhinal

Cortex

Parietal

Cortex

Entorhinal

Cortex

Anterior Dorsal

Thalamic Nuc.

Lateral

Mammillary

Nucleus

Dorsal

Tegmental

Nucleus

Postsubiculum

Taube (2007) Ann Rev Neurosci

White cue card acts as the sole intentional orienting cue.

Taube et al. (1990) J Neurosci

180° Cue Card Rotation

The cue card exerts stimulus control over

head-direction cell firing

What brain areas are important for visual landmark

control of Head Direction & Place Cell activity?

Place cell

Place cell place fields also shift following

rotation of the salient visual cue

What brain areas are important for visual landmark control

of Head Direction & Place Cell activity?

General view for processing visual information:

Dorsal stream important for processing spatial information - Parietal

cortex




Dorsal stream important for processing spatial information - Parietal

cortex

Ventral stream important for processing object recognition - Inferior

temporal cortex




Dorsal stream important for processing spatial information – Parietal

Cortex

Ventral stream important for processing object recognition – Inferior

Temporal Cortex

Visual Tectal Pathway - Attention

To test this, we conducted 90° landmark rotation experiments on HD

cells in animals with lesions of various brain areas.

Head Direction cell responses to 90° cue card

rotations

Amount of ShiftFrequency Distribution

for Shift Amounts

Control

90

270

0180

Arrow represents

the mean vector

Generative

Circuit

PoS

Entorhinal

CortexHippocampus

ADN

LMN

Subcortical Areas: Dorsal Tegmental Nuc. & Supragenual Nuc.

Head Direction Signal

Generative Circuit

PoS: Postsubiculum

ADN: Anterodorsal Thalamus

LMN: Lateral Mammillary Nuc.

PoS: Postsubiculum

ADN: Anterodorsal Thalamus

LMN: Lateral Mammillary Nuc.

Generative

Circuit

PoS

Entorhinal

CortexHippocampus

ADN

LMN

Subcortical Areas

Head Direction Signal

Generative Circuit

Control

Hippocampus

lesion

Limbic Pathway

None

Visual

Cortex

Parietal

Retspl

PoS

Entorhinal

CortexHippocampus

Visual Streams for Processing Landmark Information:

Dorsal Stream

Visual

Cortex

Parietal

Retspl

PoS

Entorhinal

CortexHippocampus


Dorsal Stream

ADN

90

270

0180

Parietal Cortex

lesion

Dorsal Stream Pathway

Retrosplenial

Cortex lesion

90

270

0180

Control

Hippocampus

lesion

Limbic Pathway

Mild-

Moderate

None

Visual

Cortex

Parietal

Retspl

PoS

Entorhinal

CortexHippocampus


Ventral Stream

Visual

Cortex

Parietal

Retspl

PoS

Entorhinal

CortexHippocampus


Ventral Stream

ADN

Entorhinal Cortex

lesion

90

270

0180

Parietal Cortex

lesion

Dorsal Stream Pathway Ventral Stream Pathway

Retrosplenial

Cortex lesion

90

270

0180

Control

Hippocampus

lesion

Limbic Pathway

None

Visual

Cortex

Parietal

PoS

LDN

Tectal Stream

Entorhinal

CortexHippocampus


Tectal Stream

via

LDN: Lateral Dorsal

Thalamus

Superior Colliculus

Pulvinar

Superior

Colliculus

Visual

Cortex

Parietal

PoS

LDN

Entorhinal

CortexHippocampus


Tectal Stream

via

LDN: Lateral Dorsal

Thalamus

Superior Colliculus

Pulvinar

Tectal Stream

Superior

Colliculus

Entorhinal Cortex

lesion

90

270

0180

Parietal Cortex

lesion

Lateral Dorsal

Thalamus lesion


Tectal

Pathway

Retrosplenial

Cortex lesion

90

270

0180

Control

Hippocampus

lesion

Limbic Pathway

None

Visual

Cortex

Parietal

PoS

LDN

Tectal Stream

Entorhinal

CortexHippocampus


Direct Projection:

Visual Cortex PoS

Postsubiculum

Visual

Cortex

Parietal

PoS

LDN

Tectal Stream

Entorhinal

CortexHippocampus


ADN

LMN

Direct Projection:

Visual Cortex PoS

Postsubiculum

Place cells

HD cells

HD cells

Entorhinal Cortex

lesion

90

270

0180

Parietal Cortex

lesion

LMN Recording

90

270

0180


Retrosplenial

Cortex lesion

90

270

0180

Control

Hippocampus

lesion

Limbic Pathway

Lateral Dorsal

Thalamus lesion

Tectal

Pathway

ADN Recording

Postsubiculum

Lesions

Hippocampal Place

Cell Recording

Place field absent

Severe

Visual

Cortex

Parietal

PoS

LDN

Tectal Stream

Entorhinal

CortexHippocampus


Direct Projection:

Visual Cortex PoS

Postsubiculum

ADN

LMN

Place cells

Entorhinal Cortex

lesion

90

270

0180

Parietal Cortex

lesion

Hippocampal Place

Cell Recording

LMN Recording

90

270

0180

Place field absent


Retrosplenial

Cortex lesion

90

270

0180

Anterior Dorsal

Thalamus lesion

Place cell recording

ADN Recording

Postsubiculum

Lesions

Control

Hippocampus

lesion

Limbic Pathway

Lateral Dorsal

Thalamus lesion

Tectal

Pathway

None

Visual

Cortex

Parietal

Retspl

PoS

LDN

Tectal Stream

Entorhinal

CortexHippocampus

ADN

LMN

Subcortical AreasGe

nera

tive

Circu

it

So, what can

we conclude ?

Superior

Colliculus

Visual

Cortex

Parietal

Retspl

PoS

LDN

Tectal Stream

Entorhinal

CortexHippocampus

ADN

LMN

Subcortical AreasGe

nera

tive

Circu

it

Landmark control in

ADN and LMN

occurs because of

the feedback loop

from PoS LMN

and ADN.

Conclusion: For

processing visual

landmark information:

the direct pathway

from Visual Areas

17, 18 --> PoS is

critical.

Superior

Colliculus

How fast can Head Direction cells learn

about the visual landmarks ?

Rats can learn landmark information very fast

Here they learned about the visual cue card in a matter of minutes.

8 min Exposure

1 min Exposure

8 min Exposure

1 min Exposure

5 out of 8 rotated well

Goodridge et al. (1998) Beh Neurosci

Is the hippocampus important for HD cells to learn

about novel visual landmarks?

Hippocampal lesion

• We can take advantage of the fact that HD cells have different preferred

directions in geometrically different environments.

• Thus, a change in the shape of the environment will usually lead to a shift in the

cell’s preferred firing direction

Cylinder vs. Rectangle & Square

Cylinder 1

Rectangle

Cylinder 2

Square

Cylinder 3

Suggests that learning the spatial information about landmarks is more like

perceptual learning, where visual information can drive the system directly – it is

not like episodic learning, which involves the hippocampus.

HD Cell preferred directions remain stable in a novel

environment across days, despite the absence of hippocampus

Cylinder – Familiar Novel Enclosures

Triangle

Pentagon

Golob & Taube (1997) PNAS

Novel Enclosures across days

What is the evidence that HD cells

can guide behavior ?

Linkage between HD cell firing and behavior

•What you would like to see is that when a cell’s preferred

firing direction shifts a certain amount, that the animal’s

spatial behavior (choice) also shifts the same amount.

•Review a few studies with HD cells where we looked for

this linkage.

•What happens with HD cells following a behavioral error?

Does the HD also show a shift in its PFD?

Training Trials Test Trial

Example 1: HD Cell information is not always used for solving

Spatial Tasks : Tolman Sunburst Maze

Muir & Taube (2004) Beh Brain Res

HD Cell Plots on Error Trials for 2 Different Cells

Training Trials Test Trials in which rat made an error

……. thus, directional heading information was present,

but the animal did not use it to guide behavior.

This could explain why they had poor performance – no learning took place.

Standard Cue Rotation Novel Environment

Example 2: Animal solves a spatial task, but is not using

HD cell information to guide its behaviorWater Reward

• Rat trained in square to go to a corner with water reward relative to

landmark.

• Receives 40-60 trials in standard session and performing ~ 77% correct.

• Rotation of landmark leads to an equal shift in the animal’s behavior and

the HD cell’s preferred firing direction.

• Rat generalizes well in the rectangle – and goes to the correct corner –

78% correct.

• But HD cells shifted their preferred directions > 72° in 12 / 13 sessions

(but their behavior did not shift the same amount) little evidence that HD

cell firing was guiding the animal’s behavior.

Behavior was not in register with the shift in the cell’s preferred direction.

La

nd

ma

rk C

ue

Square Session

Rectangle Session

Square Session

Rectangle Session

HD cell preferred direction

remains the same

HD cell preferred

direction shifts

Golob et al (2001) Beh Neurosci

• Recordings from same cell on two consecutive trials in the rectangle.

• In both cases, rat generalized well from training in the square and

selected the correct corner in the rectangle.

• Cell’s preferred direction was different on both trials – but behavior

was similar.

• Suggests that this cell’s firing was not strongly linked to the animal’s

behavior.

Head Direction cells and responses in 3D

Array of Disorientation Problems & Illusions in Space

Astronauts are frequently disoriented when working in

space (0-g) and often experience several types of illusions,

including:

• Visual Reorientation Illusion (VRIs)

• Inversion Illusion

• Extra Vehicular Activity (EVA) acrophobia

Which leads to a number of problems:

• Space motion sickness

• Poor spatial awareness – worry about emergency egress

• Inability to do work

• Flipping switches in wrong direction

• Otolith reinterpretation upon return

to a gravitational environment

HD cell responses in Vertical Plane

Experiment 1 Experiment 2

Start

Finish

Platform

Rotation

South Wall

We

st W

all

East W

all

North Wall

0°

90°

180°

270°

0°

90°

180°

270°

xy 0°

90°

180°

270°

0°

90°

180°

270°

0°

90°

180°

270°

Floor

Animal defines its horizontal reference frame as the plane it happens to

be locomoting in. Thus, it rotates its plane of locomotion by 90° as it

moves into the vertical plane and defines this new surface as its

horizontal reference frame.

South Wall

We

st W

all

Ea

st W

all

North Wall

Floor

Room Reference Frame

UpDown

Direction of Cell Firing Along Walls

HD cell responses in 3D: Vertical plane and

when Upside-down on ceiling

Calton & Taube (2005) J Neurosci

Barrier

Similar to our wall-climbing findings, HD cells

treated the walls as though they were an extension

of the floor.

Cell fires duringclimb up.

Cell fires duringclimb down.

Preferreddirectionof cell.

Running in the reverse direction, the cell is silent

on both walls.

Cell fires duringclimb up.

Cell fires duringclimb down.

Preferreddirectionof cell.

Cell is silent

during climb up.

Cell is silent during

climb down.

For most cells: Loss of directional tuning on

ceiling, with increased background firing rate

Cell 1 Cell 2

A few Cells Remained Directionally-tuned

Two examples:

…. but even most of these cells had markedly reduced peak firing rates.

During inverted locomotion on the ceiling, cells had a lower

directional information content score.

Directional Information Content

Inverted Hole Board Escape Task

Could animals learn a spatial task upside-down ?

Valerio, Clark et al. (2010) Neurobiol Learning & Memory

+

X = Four Release Points

+ = Center Release Point

x

x

Escape hole

x

x

Performance in the Inverted Hole Board Escape Task

Esca

pe

La

ten

cy (

se

c)

Training day

0

5

10

15

20

25

30

0 5 10

1 Start Point

1SP 2SP

Num

be

r o

f S

essio

ns

to C

rite

rio

n

0

4

8

12

16

Criterion Center

Probe

La

ten

cy (

sec)

0

10

20

30

Regular

trialsBlindfolded-

to-apparatus

trials

La

ten

cy (

se

c)

0

10

20

30

Surround

Curtain

Probe

Criterion

La

ten

cy (

sec)

0

10

20

30

1 Start Point

2 Start Points

Esca

pe

La

ten

cy (

se

c)

Training day

0

5

10

15

20

25

30

5 10 15 20 25

4 Start Points

30

Conclusions from Behavioral Task:

When task was simple (1 or 2 start points) – animals could

use a directional (or beacon) strategy – move toward a

particular landmark.

But when task was difficult (4 start points) – animals

needed a more flexible representation of their

environment.

They needed a flexible cognitive map-like spatial strategy.

It is possible that ‘normal’ HD activity is required for

generation and use of a cognitive map.

Valerio, Clark et al. (2010) Neurobiol Learning & Memory

+

X = Familiar Release Point 1

X = Familiar Release Point 2

+ = Center Release Pointx

x

Escape hole

Record HD cells in the

Inverted Hole Board Escape Task

Familiar

Tests

2 Start

Points .

N

S

W E

NCenter

Tests

180

S

W E

.

N

S

W E

180

S

W E

N

Sample PathsTrajectories

What are Head Direction Cells doing in this Task?

2nd Session on Floor:

Upright

Session on Floor:

Upright

Head Direction

Firin

g R

ate

Rats trained from two locations

NW

SW

Thus, rats were accurately performing the simple version of the inverted

spatial task - despite the absence of a HD signal.

NW Inverted Trial

Correct

Inverted Trial: Center

Error

SW Inverted Trial

Correct

I n v e r t e d T r i a l s

Conclusions

The results suggest that the head direction signal is

needed for accurate navigation in situations that require a

flexible representation of space (i.e., an allocentric

cognitive-mapping strategy), but not for situations that

utilize habit-like associative spatial learning.

Looking for the linkage between HD cells and

behavior using a task involving path integration.

What happens with HD cells following

a behavioral error?

Reorientation

Cells get Reset to their previously established relationship with the

familiar landmarks

I thought that building

would be on my left

Place Cell Remapping

Bostock, Kubie, Muller, 1991, Hippocampus

White

Cue Card

Black

Cue Card

Cylinder vs. Rectangle & Square

Head Direction Cell Remapping

Taube, Muller, Ranck, 1990, J Neurosci

A change in the shape of the environment usually

lead to a shift in the cell’s preferred firing direction.

In addition to Resetting, cells can also Remap

Food-Carrying task requiring Path Integration:

Behavior compared to Cell Activity

Refuge

Return

Point

Barrier

Food

Pellet

Outbound path

Return path

Valerio & Taube (2012) Nat Neurosci

Rats are blindfolded

Curtain

Hypothetical Results for Two Consecutive Trials

Initial Refuge session

(6 min)

80

0

20

40

60

Firin

g r

ate

(spik

es/s

ec)

36060 120 180 240 3000

Head direction (°)

PFD = 90°

PFD = Preferred Firing Direction



Trial 1Initial Refuge session

(6 min)

Heading error = - 30°

PFD shift = - 30°

80

0

20

40

60

Firin

g r

ate

(spik

es/s

ec)

36060 120 180 240 3000

Head direction (°)

PFD = 90°80

0

20

40

60

36060 120 180 240 3000

PFD = 60°

Head direction (°)

- 30°



Trial 1Initial Refuge session

(6 min)

Refuge

Inter-trial 1


PFD shift = - 30°

Reorientation:

Resetting

80

0

20

40

60

Firin

g r

ate

(spik

es/s

ec)

36060 120 180 240 3000

Head direction (°)

PFD = 90°80

0

20

40

60

36060 120 180 240 3000

PFD = 90°

Head direction (°)

80

0

20

40

60

36060 120 180 240 3000

PFD = 60°

Head direction (°)

The PFD gets reset to

the previous refuge

value.

- 30°


Trial 1 Trial 2Initial Refuge session

(6 min)

Refuge

Inter-trial 1

- 30°


PFD shift = - 30°

Reorientation:

Resetting

80

0

20

40

60

Firin

g r

ate

(spik

es/s

ec)

36060 120 180 240 3000

Head direction (°)

PFD = 90°80

0

20

40

60

36060 120 180 240 3000

PFD = 90°

Head direction (°)

80

0

20

40

60

36060 120 180 240 3000

PFD = 60°

Head direction (°)

Heading error = + 60°

PFD shift = + 60°

+ 60°

80

0

20

40

60

36060 120 180 240 3000

PFD = 150°

Head direction (°)


the previous refuge

value.PFD = Preferred Firing Direction



Trial 1 Trial 2Initial Refuge session

(6 min)

Refuge

Inter-trial 1

Refuge

Inter-trial 2


PFD shift = - 30°

Reorientation:

ResettingReorientation

Correction Process:

Remapping

80

0

20

40

60

Firin

g r

ate

(spik

es/s

ec)

36060 120 180 240 3000

Head direction (°)

PFD = 90°80

0

20

40

60

36060 120 180 240 3000

PFD = 90°

Head direction (°)

80

0

20

40

60

36060 120 180 240 3000

PFD = 60°

Head direction (°)

Heading error = + 60°

PFD shift = + 60°

80

0

20

40

60

36060 120 180 240 3000

PFD = 150°

Head direction (°)

80

0

20

40

60

36060 120 180 240 3000

PFD = 150°

Head direction (°)


the previous refuge

value.

The PFD does not get

reset to the previous

refuge value.

- 30°

+ 60°

Individual Trials for 3 Different Cells

Behavioral Heading Error correlates with Shift of HD cell’s PFD

Inter-trial interval in refuge

During foraging task

Heading Error = - 7.5°

PFD Shift = - 8.6°

Heading Error = + 8.5°

PFD Shift = + 18.0°


PFD Shift = - 36.8°

Head Direction (deg) Head Direction (deg) Head Direction (deg)

600 120 180 240 300 360 600 120 180 240 300 360 600 120 180 240 300 360

60

40

20

0

Firin

g R

ate

40

20

0

30

10Firin

g R

ate

40

20

0

30

10Firin

g R

ate

Return trip

Correlation between Heading Error and the amount the

Preferred Firing Direction (PFD) shifted during the

outbound trip was good

Headin

g E

rror

(°)

r = 0.72

PFD shift from refuge to when

rat found food pellet

180

90

0

-90

-180

180900-90-180

Example When Resetting Process Occurred

Heading Error = 5.7°

PFD Shift = -5.3°

0

10

20

30

40

3600 60 120 180 240 300

Head Direction (°)

Firin

g R

ate

Outbound path

Return path

Original

Refuge

PFD = 102°

Outbound Response

Original (initial) Refuge Response

Trial 1



PFD Shift = -5.3°

0

10

20

30

40

3600 60 120 180 240 300

Head Direction (°)

Refuge

Inter-trial 1

0

10

20

30

40

3600 60 120 180 240 300

Head Direction (°)

Firin

g R

ate

Outbound path

Return path

Original Refuge Response

Current Refuge Response

Original

Refuge

PFD = 102°

Return

Refuge

PFD = 108°

Trial 1



PFD Shift = -5.3°

0

10

20

30

40

3600 60 120 180 240 300

Head Direction (°)


PFD Shift = -22.3°

Refuge

Inter-trial 1

0

10

20

30

40

3600 60 120 180 240 300

Head Direction (°)

Firin

g R

ate

0

10

20

30

40

3600 60 120 180 240 300

Head Direction (°)

Error

Outbound path

Return path


Preceding Refuge Response

Outbound Response

Original

Refuge

PFD = 102°

Return

Refuge

PFD = 108°Foraging

PFD = 80°

Trial 1 Trial 2



PFD Shift = -5.3°

0

10

20

30

40

3600 60 120 180 240 300

Head Direction (°)

0

10

20

30

40

3600 60 120 180 240 300

Head Direction (°)


PFD Shift = -22.3°

Refuge

Inter-trial 1

Refuge

Inter-trial 2

0

10

20

30

40

3600 60 120 180 240 300

Head Direction (°)

Firin

g R

ate

The PFD is reset to the refuge value upon return to the refuge.

0

10

20

30

40

3600 60 120 180 240 300

Head Direction (°)

Error PFD is Reset

Outbound path

Return path



Original

Refuge

PFD = 102°

Return

Refuge

PFD = 108°Foraging

PFD = 80°Refuge

PFD = 105°

Trial 1 Trial 2

Example When Remapping Process Occurred

Outbound path

Return path

Head Direction (°)

3600 60 120 180 240 300

80

0

20

40

60

Firin

g R

ate



Original

Refuge

PFD = 196°

Refuge Response

Outbound Response

Trial 1


Outbound path

Return path

Head Direction (°)

3600 60 120 180 240 300

80

0

20

40

60

Refuge

Inter-trial 1

Head Direction (°)

3600 60 120 180 240 300

80

0

20

40

60

Firin

g R

ate



PFD is not Reset



Original

Refuge

PFD = 196°

Return

Refuge

PFD = 125°

Trial 1


Outbound path

Return path

Head Direction (°)

3600 60 120 180 240 300

80

0

20

40

60

Refuge

Inter-trial 1

Head Direction (°)

3600 60 120 180 240 300

80

0

20

40

60

Firin

g R

ate


PFD Shift =5.5°



PFD is not Reset

Head Direction (°)

3600 60 120 180 240 300

80

0

20

40

60


Preceding Refuge Response

Outbound Response

Original

Refuge

PFD = 196°

Return

Refuge

PFD = 125°Outbound

PFD = 123°

Trial 1 Trial 2


Remapping Occurs - the PFD is not reset to the refuge value upon return

to the refuge, but instead retains the current PFD value (~120°) for the

next trial (#2) and subsequent trials (#s 8 and 9 are shown).

Outbound path

Return path

Head Direction (°)

3600 60 120 180 240 300

80

0

20

40

60

Head Direction (°)

3600 60 120 180 240 300

80

0

20

40

60

Refuge

Inter-trial 1

Refuge

Inter-trial 2

Head Direction (°)

3600 60 120 180 240 300

80

0

20

40

60

Firin

g R

ate


PFD Shift =5.5°



PFD is not Reset

Head Direction (°)

3600 60 120 180 240 300

80

0

20

40

60

Original

Refuge

PFD = 196°

Return

Refuge

PFD = 125°Outbound

PFD = 123°

Return Refuge

PFD 2 = 118°PFD 8 = 113°PFD 9 = 136°

Inter-trial 8

Inter-trial 9

Inter-trial 2Original Refuge

Trial 1 Trial 2

When did Resetting and Remapping trials occur?

12090 1500 30 60

Resetting trials

Remapping trials

Heading Error (°)

Nu

mbe

r of

tria

ls

Resetting occurred more often following small heading errors.

Remapping occurred more often following large heading errors.

Conclusions

•HD cells encode information about the animal’s perceived directional

heading in its environment.

•Can learn landmarks quickly and without the hippocampus.

•HD cells do not appear to be used when animal is using a directional

‘beacon’ strategy, but may be necessary for spatial strategies that require a

flexible representation of the spatial environment - cognitive map.

•HD cell activity and behavioral performance correlated well when animal

performed a task requiring path integration.

•When a behavioral error is made, there are two types of correction

processes: resetting and remapping.

Grid Cell Signal Generation

Grid Cells & Motor/Proprioceptive Cues

learning & memory in the head direction cell...

Documents