Natural Computation and Self Organization…in Monkey Behavior

NaturalComputationandSelfOrganization…inMonkeyBehaviorKELLYFINN

ANIMALBEHAVIORGRADUATEGROUP

Overview

- commonmethodsinanimalbehavior

- myinterests

- rationale

- goals/aims

- preliminarydata

- preliminaryanalysis

TraditionalSamplingObservationalmethods(Altmann,1974;Martin&Bateson,2007)

Ethogram– adetailedcatalogueofbehaviors,usuallymutuallyexclusiveandasobjectiveaspossible.Ideally,itshouldbeabletoinstructtwodifferentpeopletorecordthesamebehaviorsthesameway

Typesofsampling:

- Alloccurrence

- 1/0sampling

- Instantaneousscansampling

Mostlyuseintuitiontodecideuponsampling,timescales,durations.

Newtoolsandtechnology

MyInterests– patternandstructure- Howtomeasurebehavior

- Howpatternsrelatestofunction

- Howanimalsperceiveandrespondtopatterns

Dissertation:

1.Patternperception

2.Individualbehaviorpatterns

3.Grouplevelnetworkstructure

Systemisananimal– itsbehavior isaproductofmanydifferent thingsonmanydifferenttimescales:evolutionaryhistoryofthespecies,geneticcomposition, rearinginfluences,interactionswithotheranimals,currentenvironment,pastenvironment, hormones, healthstate,physiology, etc.etc.etc.(andallof thesethingsalsointeractwitheachother).

Ananimalisbasicallyablackbox,withbehavioraloutputswecanmeasure.

Complexityinbiologicalsystems

‘Theoryofcomplexity lossinaginganddisease’ (Goldbergeretal,2002)- healthysystemsoutputcomplexmulti-scaled variabilitythatallowsforadaptivecapabilities andbreaksdownwithageorillness.- the long-rangescalingpatternsofhumanheartbeatandgaitdynamicshavebeenshowntodeteriorateinmarkedlypathologicalconditions(Goldbergeretal,2002b).

“Acompletelyordereduniverse,however,wouldbedead.Chaosisnecessaryforlife.Behaviouraldiversity,totakeanexample,isfundamentaltoanorganism’ssurvival.Noorganismcanmodeltheenvironmentinitsentirety.Approximationbecomesessentialtoanysystemwithfiniteresources.Chaos,aswenowunderstandit,isthedynamicalmechanismbywhichnaturedevelopsconstrainedandusefulrandomness.Fromitfollowdiversityandtheabilitytoanticipatetheuncertainfuture.Thereisatendency,whoselawswearebeginningtocomprehend,fornaturalsystemstobalanceorderandchaos,tomovetotheinterfacebetweenpredictabilityanduncertainty.Theresultisincreasedstructuralcomplexity.”– Crutchfield(2011)

Complexityinbehaviorpatterns

- animal-environmentinteraction(i.e.resourceacquisition)

- self-regulationofbehaviorinresponsetoachangingenvironment

- pathologicalbehavioralstereotypy

Themeasurementofbehavioralcomplexityorstructuremaybeanuntappedyetfundamentalsourceofinformationaboutananimal’sbehavioralecologyandhealth.

BriefreviewofliteraturePreviousstudieshavereportedthatcertainanalysesofbehaviorpatterncandetectsubtlechangesin‘hidden’animalbehavioralstructurethataremissedwhenusingtraditionalmeasuressuchasaveragedurations(Escós etal.,1995;Asheretal.,2009;Seuront &Cribb,2011).

Variabilityinbehavioralsequencecomplexityhasbeenlinkedtomanycharacteristics:

- differencesbysex:socialbehaviorofchimpanzees(Alados &Huffman,2000);foragingandlocomotionbehaviorinJapanesemacaques(MacIntosh etal.,2011)

- stress level:boattrafficdisturbance indolphinsurfacebehavior(Seuront &Cribb,2011);resourcedeprivationinchickens(María etal.,2009)

- age:youngchickens(María etal.,2009);adultmacaques(MacIntosh etal.,2011)

- statesknowntobeenergeticallytaxingsuchasreproduction,clinically impairedhealth,andparasiteinfection:Japanesemacaques(MacIntosh etal.,2011),SpanishIbex(Alados etal,1996),chimpanzees(Alados &Huffman,2000).

AnimalMovement

- fairlyobjectivebehaviortoscore

- fundamentaltosurvivalofmobilespecies,multi-purpose

- alreadyassociatedwithhealth(self-tracking)

- tonsofthistypeofdataisbeingcollected

“Movement data provide a window - often our only window -into the cognitive, social, and biological processes that underlie the behavioral ecology of animals in the wild.” - Gurarie et al 2016

Goals:

Describestructureinmonkeymovement

- Relatepatternattributestofunction◦ behavioralstate,terrain,individual characteristics(age,health,rank,etc.)

- Informfuturemeasurementandsampling

Koshima,Summer2015- EAPSIprogram(NSF/JSPS)

- Pilotdata

- collaborationwithDr.AndrewMacIntosh

- behavioralstructure~parasiteload

JapanesemacaquesofKoshima

Lifeinthefield

http://mandelbrotmonkey.blogspot.com

Data

- 56sequences(4sequencesfromeachof14monkeys)

- 1-0activitydowntothesecond*

- 1,000-4,000datapoints/sequence

- Activityrangedfrom.5%to52%(mean17.51%)

0.00

0.02

0.04

0.06

0 200 400d1$length

density

typeact

inact

"density plots of lengths of activity bouts"

Sequences

ReturnMaps

0 20 40 60 80 100 120 140

0100

200

300

400

500

Activity followed by inactivity bout lengths

ilength alength

ilength

0 20 40 60 80 100 120 140

020

4060

80100

120

140

Activity bout lengths n+1~n

n

n.1

0 100 200 300 400 500

0100

200

300

400

500

Inactivity bout lengths n+1~n

n

n.1

H(L)fordifferentL,usingwordfrequencies

o

o

o

o

o

o

o

o

o

2 4 6 8

0.5

1.0

1.5

2.0

2.5

HxL for different monkey sequences

d$lcb

ind(

d$so

, d$s

k, d

$sku

, d$s

s, d

$sso

, d$s

u)

oo

oo

oo

oo

o

oo

oo

oo

oo

o

oo

oo

oo

oo

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

(Goldenmean)

random random bout real

2

4

6

8

0.0 2.5 5.0 7.5 0.0 2.5 5.0 7.5 0.0 2.5 5.0 7.5

word length

entropy

monkeykibana

omoto

sone

usu

Entropy by word length for randomized and real sequences

Nextsteps- E-machines◦ BayesianInferencetogetinfocalculations◦ Alternating renewalprocessmodeltype

Marzen etal,2015

Staytuned…

- FieldseasonWinter2017- Partitioningfromaccelerometerdata- Simulateddata,sampling