Studies of Neural Time from Neuroethology.     PL 393     Prof. J.D. Pettigrew

What is Neuroethology?

Ethology: Study of natural behaviour: 3 Nobel Prizes (1973)
Karl Von Frisch.....bee dance
Konrad Lorenz...........filial imprinting
Niko Tinbergen.......innate release mechanisms

Neuroethology: Natural behaviour is analysed with all the modern methods of neuroscience. A striking behaviour observed in natural conditions can then be understood in neural terms, and vice versa

Examples of Neuroethological Systems:
Konishi..........auditory localisation by barn owl, song learning in passerine birds
Nottebohm.......song learning nuclei of the avian brain
Heiligenberg.....jamming avoidance response of electric fish
Suga..............auditory processing by echolocating microbats
Wehner..........navigation by the desert ant , Cataglyphis.
Pettigrew.........binocular vision in mammals, birds, fish etc
Land...............eye movement and optical strategies in vision
Menzel...........colour vision in bees
etc etc
 

Behaviour guides neural analysis: Note that  the neural analysis often requires insights from the natural behaviour. In the lecture this will be emphasised with respect to Sugaís work, which could not have been undertaken without a full knowledge of the natural behaviour because the extremely-specific stimulus conditions (e.g. FM1 FM2 FM3 followed by echoes of the same tones with a delay) could never be discovered by an investigator without detailed knowledge of the natural behavior (in fact, some of teh full richness of teh neural behaviour was missed at first because the apparatus was not precise or specialised enough).

Neural Analysis Illuminates Behaviour: The process also works in the reverse direction, because the detailed information made available from brain recordings (the animal can effectively ìtellî the investigator its preferences etc) can illuminate unsolved puzzles in behaviour. The example that I will use in class is the puzzle of how delays of returning echoes are measured by microbats. What is the "initiating signal" from which counting begins to measure the delay? A reafferent signal from the motor system to the auditory system that signifies that a sonar signal is sent? Or does the count begin when the bat hears its own sonar signal? Neuroethology has provided the likely answer.

Neural Representations of Time:

1. Feynman's Unanswered Question:
Does the brain have a fast clock to synchronise its many related activities in the same way a computer is strobed by its fundamental clock?

New interest in the role of brain rhythms to 'bind' image segments (W. Singer) or to imprint odour patterns (W. Freeman) has kept Feyman's question currrent, if unanswered.

2. Time vs. Space Choreography:
There is a great need for more kowledge about neural time to complement the increasing information about neural space. e.g. homeobox genes provide beautiful examples of the production of spatially-repetitive structures, but temporal periodicity in development (just as important if patterns are to be generated) is only just being studied (e.g. cHairy and lunatic fringe genes that have 90 min periodicity in the production of somites in chick embryos).

3. Representation of neural time does not map onto experiential time.
Note the special problems here, where a temporal pattern has to be represented with neurons.

cf. 1. Penfield's experiments with electrical stimulation of the temporal lobe of humans undergoing neurosurgery; the impression can be MUCH longer than teh duration of the stimulation.
2. 'Filling in' of a long dream sequence that was very short-lived; again, the perception of duration bears no relation to teh duration of the related mental activity.

Introspecting Two Clocks:
Experiential Timing: Imagine two clocks, one running faster than the other. Imagine you are the clock, aware only of intervals between ticks. Since the slow clock will have a larger ëslabsí of ëexperienceí that intervene between each of its ticks, more ëexperienceí will pass it by. Paradoxically then, the slower ticking clock will experience more rapidly passing time.

This viewpoint helps to explain those moments when time appears to slow down, or even stop.
e.g. as you are hurtling toward an imminent collision; or when you suddenly meet a very special person after a long absence.

The slowing of time that one experiences in those situations is a function of the 'speed-up' of the clock rate of key brainstem nuclei like the locus coeruleus.

On the other hand, the whole thing can be turned around if one is conscious of the passage of time and goes about marking its passage. This would be as if the two clocks each had eyes that they could use to look around and compare their rates with each other. In this situation, the internal feeling that the faster clock was experiencing a slower passage of time would be overidden by the 'count' of intervals of the clock's hands that was observed 'directly from outside'.

Interval Timing: In this latter case, the timing mechanism is called ëinterval countingí. This is the mechanism one uses to wake at a specific time, or to estimate the passsage of time accurately. Very accurate interval timing is important in many species. e.g. small birds foraging in the Arctic at the approach of winter have to know exactly how to budget time in different places (Melissa Bateson).

Interval timing takes place in the basal ganglia, e.g. substantia nigra, according to the results of brain scanning studies. Time distortions are a common by-product of the effects of dopaminergic drugs that act on the basal ganglia.

These two different time systems in the brain, experiential timing and interval timing (there are bound to be more varieties of time, such as circadian time) help to account for the paradoxical differences we may experience in time.
 ëTime stood still when they metí
cf. ëTime flies when you are having funí
[The second aphorism stands in contrast to the first example, where slowing of time occurs during intense excitement, (rapid ticking of the 'clock') such as with an imminent collision or an intensely-exciting rendezvous. The difference probably can be accounted for by the 'self-conscious' element that brings in interval counting....just as a boring wait can seem interminable once we become conscious of it and start counting the moments].

Measuring Small Time Intervals:
Although communication around the nervous system utilises action potentials that last around 1 millisecond, much smaller time intervals are regularly measured in a variety of time-as-place structures.

1. Electric fish:
Mormyrids (African) and Gymnotiforms (S. American) are 'weakly electric fish. Some of their many species hold the record for time discrimination in vertebrates (microbats and marine mammals using sonar come a close second). The jamming-avoidance response of Gymnotus (Mormyriforms) can detect 10 nanoseconds. The aggregations of neurons involved (e.g. cerebellum and electroreceptive nucleus) are enormously enlarged, in keeping with ideas that a larger array of neurons can divide up the time into a larer number of smaller intervals.

2. Microbat audition:
Simmons has examples of microbat discriminations that appear to involve time intervals as small as the electric fish, around 10 ns, although not everyone agrees with the interpretation.
Combination sensitive neurons in Pteronotus auditory cortex recorded by Suga and his colleagues, have indisputable optimal time intervals in the microsecond range.
The importance of this work lies in the mechanism of ìinitialisingî the time measurement. By looking at the varieties of 'combination sensitive' neurons that measure delay between the outgoing sonar signal and teh returning echo, one can see that some are common and some are unknown, even though physically possible. In particular, one never sees combination- sensitive neurons that are sensitive to a returning first harmonic (FM1). In other words, while one sees FM1-FM2, FM1-FM3, FM2-FM2, FM3-FM3......one does not see FM1-FM1, FM2-FM1 or FM3-FM1. Since the first harmonic is often suppressed by microbats, the most parsimonious explanation of these findings is that the bat "self-stimulates" with its own call to initiate the delay measuring process and that the first harmonic is the most important "self-stimulus". By adjusting the intensity of its first harmonic so it is weak compared to the other very intense components of the call, the bat can ensure that the returning echo of the first harmonic is inaudible. In this way confusion can be reduced.
    From the point of view of the present question, How is the delay measurement initialised?....it is hard to think of any other explanation for the distribution of types of combination sensitive neurons, except that they reflect a system that initialises by self-stimulation. Suga's group have verified this conclusion with a series of experiments that modify self-stimulation.
 

For example, Sugaís work on the delay-sensitive neurons in auditory cortex of the moustache bat has made it clear how the bat ìinitialisesî the time counter that measures the delay from the emission of the sonar pulse to the return of the echo from the prey object. The ìinitialisationî is triggered by self-stimulation caused by FM1 in the outgoing pulse (rather than by a signal from the motor system, for example). By adjusting he intensity of FM1, the bat can use this for self-stimulation and then use the echo from the louder FM2 or FM3 to measure the delay.
 
 

Microbat Auditory Cortex:  Suga
Ref: Scientific Amer.June 1990 Vol 262 #2 pp 60-66 ìBiosonar and neural computation in batsî
CF-FM bats with Doppler shift compensation:
Allows detection of moths with beating wings in acoustic clutter.

25 years work on he auditory cortex has revealed some details of the ìcontrol panel in the batís cockpitî (the multiple auditory cortical areas specialised for different functions).
1. Doppler-shift area (central representation of the acoustic fovea).
2. FM-FM combination sensitive, Delay-tuned area: Target Distance
3. CF-CF combination sensitive, Target Velocity area
4. Target Azimuth Area  etc etc

The oganisation of the FM-FM Delay Sensitive Area revelas the strategy used by the bat to measure delay of the echo. Only 3 combinations are represented, FM1-FM2, FM1-FM3 and FM1-FM4, revealing that the bat is using FM1 as the initialising signal. Since FM1 is weak and can be adjusted in intensity, the bat can use this tone to self-stimulate its auditory system on sonar pulse emission and then ìlistenî (via the FM-FM neurons) to the returning echo of FM2 or FM3 or FM4. The FM-FM neurons in the Delay Tuned area can then read out the delay of the echo (target distance) . This arrangement avoids the need for the motor system to communicate the time of sonar emission to the auditory system, and minimises confusion that could arise from the sonar of other bats, since the quiet FM1 is just loud enough to be heard by the bat itslef, but not loud enough to carry far or produce an echo from a distant target.
 

3. Motor Cortex: Throwing Spears and Spearing Semantics (Words)
Bill Calvin has argued,  in a number of readable books on the brain, that an increase in temporal precision is one of the driving forces behind the expansion of hominid cortex. Extreme temporal precision is required for accurate throwing and this may have helped the evolution of asymmetry and enlargement of the forebrain.
Considerable temporal precision is involved in word production, formants at the appropriate intervals by the Left hemisphere which has good temporal precision, in contrast to the prosodic variations in tone produced by the Right hemisphere. This is one good reason for the R/L asymmetry observed in hominid temporal cortex........more neurons being required, a la electric fish, to come up with the required temporal precision.

Oscillators and Clocks:
Molecular details of neural clocks are gradually becoming available. There is still no compelling picture of how the rhythms are generated by any of these oscillators (I personally find it unsatisfying, for instance, that there is no idea where one obtains a time constant of around 24 hours from a bunch of proteins like TIM and PER. Teresa Chay has some models that are much more attractive and plausible to me than all the molecular biology of per and tim. Teresa focusses on calcium ion fluxes into the major intracellular compartments, such as the endoplasmic reticulum. It is easy to imagine the ER like an hour glass, gradually filling up with calcium ions, then emptying again when it fills up and reverses the polarity of ion channels)

A. Nanoseconds:
The record on the planet for short time interval detection is 10 nanosec...by Gymnotus, a mormyrid electric fish from Africa. It is not clear if there is an oscillator that mimicks this short period, but the mechanism is generally believed to be the considerable enlargement of the neural apparatus  for electroreception, especially the torus and cerebellum that are enormously enlarged in these electric fish.
A delay line with more neurons strung along it can obviously divide time into smaller and smaller intervals.

B. Microseconds:
Mammals have a voltage-sensitive, Ca2+ independent, contractile protein,  in the cell wall of the motile outer hair cells of the cochlea.
This protein changes the length of the outer hair cell in synchrony with microphonic potentials in the cochlea (note the uniqely-large generating potential of +150mV).
This protein  can therefore follow oscillations at least to 100kHz (10 microsecs), the frequencies used by some marine mammals and some microbats. The invention of this new protein has enabled mammals to invade some niches that are barred to other vertebrates (like birds, whose hearing is limited to frequencies below 10 kHz--100 microsecs).

C. Seconds:
i.  Hyperpolarisation-sensitive cationic channels (Ih....also called If ...for "funny") are now described and cloned from oscillating tissues (e.g. sperm head, cardiac muscle, brain).
Triggered by hyperpolarisation, they allow a ramp of depolarisation to drift the cell's membrane potential back toward resting before shutting off. The steepness of the ramp of depolarisation determines the frequency of the oscillation, which is around the 1 sec range for Ih in the tissues described.
ii.  Bacterial Chemotaxis:
There is a 1-2 sec switch between the two phases of bacterial taxis:-   'go' (when the bacterial 'outboard motor' is spinning and propelling the bacterium in a straight line).....and 'stop' (when the flagellar power is turned off and the bacterium tumbles randomly at the mercy of Brownian motion).
The molecular details of this switch have been investigated by Dennis Bray and involve a network of proteins such as tyrosine kinases. The switch is beautifully regulated. The 'go' phase is lengthened if an attractant is detected, whereas the 'stop' phase is lenthened if a relllent substance is detected. Despite these sensory modifications of the switch period, it always returns to the baseline period, a form of adaptation.

I think that the basic machinery for the 'stop' and 'go' switch in bacteria may have been retained during evolution. It is such a fundamental aspect of behaviour and is linked so strongly to sensory processing. Whether it shares a molecular link with bacteria or not, the Right hemisphere=stop and Left hemisphere=go connection is a useful mnemonic.

D. Hours to Days:
Circadian rhythms have evolved independently a number of times, judging by the duplication of biochemical effort in different organisms.

i. Drosophila, Mammals and Birds: transcription of per and tim genes lead to translation of PER and TIM proteins which dimerise. The heterodimer turns off transcription.
ii. Neuropora. frq gene transcription and translation leads to protein FRQ turning off transcription when it gets into the nucleus, just like the per tim system of animals, but note that the genes are independent..
iii Cyanobacteria (Synechococcus). A 3 gene, negative feedback sytem kaiA, kaiBC.

The point of comparing these diverse systems is to show that all involve a transcriptional clock with negative feedback from the gene products back to the genes. A positive-feedback limb has also been found in each of these systems that re-activates the cycle.
When it comes to very slow oscillators, transcriptional clocks seem to be the way to go. ?Perhaps there were RNA clocks in the pre-DNA world.

Where does the slow speed of the circadian clock come from?
This question is still unanswered. My early favourite, the transcriptional process itself, seems to be ruled out by the fact that transcription, as slow as it is, and as large as some of these genes are (per is hundreds of kB long), is still completed for teh circadian clock genes in a fraction of the circadian period.
Current focus seems to be on the availability of the TIM and PER proteins, which are unstable and kept in low concentrations by degradative processes and membrane recycling/ubiquitination unless there is no light.
But note that pure transciptional clocks may exist without any need for protein intermediation. e.g. the 90 min period of lunatic fringe and cHairy, involved in somite formation in chick development, seems not to involve a protein translation intermediate and may be purely transcriptional.
 

References and Notes:

Outer Hair Cell Motor:
Gale JE, et al.

The outer hair cell motor in membrane patches.
Pflugers Arch. 1997 Jul;434(3):267-71.

Ashmore J, et al.

Hearing in the fast lane.
Curr Biol. 1999 Aug 29 July/12;9(15):R572-R574.

2 : Geleoc GS, et al.
A sugar transporter as a candidate for the outer hair cell motor.
Nat Neurosci. 1999 Aug;2(8):713-9.
 

 Bacterial Chemotaxis:

1 : Abouhamad WN, et al.
Computer-aided resolution of an experimental paradox in bacterial chemotaxis.
J Bacteriol. 1998 Aug;180(15):3757-64.

2 : Bray D.
Signaling complexes: biophysical constraints on intracellular communication.
Annu Rev Biophys Biomol Struct. 1998;27:59-75. Review.

3 : Bray D, et al.
Receptor clustering as a cellular mechanism to control sensitivity.
Nature. 1998 May 7;393(6680):85-8.
 

Hyperpolarisation-activated cation channels:

Biel M, et al.
Hyperpolarization-activated cation channels: a multi-gene family.
Rev Physiol Biochem Pharmacol. 1999;136:165-81. Review. No abstract available.
 
 
 

'Lunatic Fringe': 90 min oscillator coupled to somite formation:

 1. Dev Biol 1999 Mar 1;207(1):49-61

Dynamic expression of lunatic fringe suggests a link between notch signaling
and an autonomous cellular oscillator driving somite segmentation.
Aulehla A, Johnson RL

Department of Biochemistry and Molecular Biology, University of Texas, M.D. Anderson Cancer
Center, Houston, Texas, 77030, USA.

The metameric organization of the vertebrate trunk is a characteristic feature of all members of this
phylum. The origin of this metamerism can be traced to the division of paraxial mesoderm into
individual units, termed somites, during embryonic development. Despite the identification of
somites as the first overt sign of segmentation in vertebrates well over 100 years ago, the
mechanism(s) underlying somite formation remain poorly understood. Recently, however, several
genes have been identified which play prominent roles in orchestrating segmentation, including the
novel secreted factor lunatic fringe. To gain further insight into the mechanism by which lunatic
fringe controls somite development, we have conducted a thorough analysis of lunatic fringe
expression in the unsegmented paraxial mesoderm of chick embryos. Here we report that lunatic
fringe is expressed predominantly in somite -II, where somite I corresponds to the most recently
formed somite and somite -I corresponds to the group of cells which will form the next somite. In
addition, we show that lunatic fringe is expressed in a highly dynamic manner in the chick
segmental plate prior to somite formation and that lunatic fringe expression cycles autonomously
with a periodicity of somite formation. Moreover, the murine ortholog of lunatic fringe undergoes a
similar cycling expression pattern in the presomitic mesoderm of somite stage mouse embryos.
The demonstration of a dynamic periodic expression pattern suggests that lunatic fringe may
function to integrate notch signaling to a cellular oscillator controlling somite segmentation.
 

2. Curr Biol 1998 Aug 27;8(17):979-82

  The lunatic fringe gene is a target of the molecular clock linked to somite
segmentation in avian embryos.

McGrew MJ, Dale JK, Fraboulet S, Pourquie O

 Laboratoire de genetique et de physiologie du developpement (LGPD), Developmental Biology
 Institute of Marseille (IBDM), CNRS-INSERM-Universite de la mediterranee-AP de Marseille,
 France.

 The most obvious segments of the vertebrate embryo are the trunk mesodermal somites which give
rise to the segmented vertebral column and the skeletal muscles of the body. Mechanistic insights
 into vertebrate somitogenesis have recently been gained from observations of rhythmic expression
of the avian hairy-related gene (c-hairy1) in chick presomitic mesoderm (PSM), suggesting the
existence of a molecular clock linked to somite segmentation ([1]; reviewed in [2]). Here, we show
that lunatic Fringe (IFng), a vertebrate homolog of the Drosophila Fringe gene, is also expressed
rhythmically in PSM. The PSM expression of IFng was observed as coordinated pulses of mRNA
resembling the expression of c-hairy1. We show that c-hairy1 and IFng expression in the PSM are
coincident, indicating that both genes are responding to the same segmentation clock. The genes
were found to differ in their regulation, however; in contrast to c-hairy1, IFng mRNA oscillations
required continued protein synthesis, suggesting that IFng could be acting downstream of c-hairy1
 in the clock mechanism. In Drosophila, Fringe has been shown to play a role in modulating
Notch-Delta signalling [3,4], a pathway which in vertebrates has been implicated in defining somite
boundaries [5-9]. These observations place the segmentation clock upstream of the Notch-Delta
pathway during vertebrate somitogenesis.

PER and TIMELESS: Circadian Clock Genes.

: Naidoo N, et al.
A role for the proteasome in the light response of the timeless clock protein.
Science. 1999 Sep 10;285(5434):1737-41.

2 : Sangoram AM, et al.
Mammalian circadian autoregulatory loop: a timeless ortholog and mPer1 interact and negatively
regulate CLOCK-BMAL1-induced transcription.
 Neuron. 1998 Nov;21(5):1101-13.

3. Liu K Heintzen C. Loros J and Dunlap JC 1999 Cellular and Molecular Life Sciences 55: 1195-1205
Regulation of clock genes.