motion - DYNAMIC COMMUNICATION SYSTEMS Karl Grammer

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: DYNAMIC COMMUNICATION SYSTEMS Karl Grammer LUDWIG-BOLTZMANN-INSTITUTE FOR URBANETHOLOGY ..."from the invisible atom to the celestial body lost in space. Everything is movement...it is the most apparent characteristic of life: it manifests itself in all its functions. It is even the essence of several of them"... Etienne Jules Marey (1839-1904) DEBRECEN 2003 1 SYSTEM CONSTRAINTS IN INTERACTIONS u u u u High potential costs Reactive iterative solutions Minimizing costs Predictive systems Interactive scripts Mindreading Good-move theorem Avoiding costs The problem of deception Search for honest signals Biological system contraints Structures for prediction A X B X A B B A B Y Y Z DEBRECEN 2003 SOCIAL MODULAR MIND u u u u u Adaptive information processing Adaptive decision making Swiss Army Knife Theory Choosing stragies and planning actions Explosion of calculation time and complexity Knowledge aquisition which reduces search space for the selection of alternatives System stabilization Selection of Modules - Plans - Actions A Z DEBRECEN 2003 2 AFFECT: THE SUPERUSER u u Affective information dominates cognitive information Mindreading assessment of internal states non-falsifiable signals Exploitation manipulation of internal states induction of affect Affect is the prerequisite for cognitive functioning Situation approriate selection of actions Search for honest unfalsifiable affective information which is useful for decision structuring DEBRECEN 2003 MOTION AND INTERNAL STATES u u u u Laws of physics invariant thermodynamics environment of adaptedness Modulation by configuration of biomechanical linkages genetically determined Neuronal motor control strategies genetically determined variant learning Neurotransmitters dopamine and serotonine sex hormone dependend Perception of quality of motion as unfalsifiable signals DEBRECEN 2003 3 SEQUENTIAL u u u u COMMUNICATION u Information processing approach Signal frame and form Signal + knowledge = meaning Communication models encoding and decoding error control Feedback Sequential, discrete and "binary" INFORMATION INFORMATION CHANNEL SIGNAL SIGNAL SIGNAL RECEIVER RECEIVER RECEIVER SIGNAL CHANNEL CHANNEL MESSAGE MEANING DEBRECEN 2003 CLASSICAL METHODS u u Ad-lib observation Construction of repertoire categorization discreteness homogeneity stereotypy validation of repertoire concepts about behavior variations of the same signal ? empirical methods coding effort ? Categories are theories DEBRECEN 2003 4 TOO MANY MEANINGS u u u u u u u u u u modulated meaning age, sex, attractiveness time-specific meaning static-dynamic meaning parallel signals signal combinations tonic nature of signals 10 out of 50: 1010 combinations quality cultural conventions atomization one code for every behavior DEBRECEN 2003 INFERENTIAL u u u u COMMUNICATION Neurophysiological pathway F5 neurons - ventral premotoric region Movement mirroring Completion of partially seen movements DEBRECEN 2003 5 THE SHARED MANIFOLD Early imitation A Smile is not a smile Direct and real time induction of emotionsattitudes-moods DEBRECEN 2003 MOVING DOTS u u Gunnar Johansson 1964 point light displays sex emotion intention Anna Tagiuri moving dots and personality DEBRECEN 2003 6 MOVEMENT RESEARCH: NEW APPROACHES u Guido Kempter und Siegfried Frey moving politicians moving avatars personality encoded in movement Gary Bente und Nicole Kraemer perception of social interaction shift of attention social perception identical DEBRECEN 2003 u u u Motion capturing methods highly intrusive Empirical coding time-consuming Motion Energy Detection Model free descriptions Simple implementation MOTION CAPTURING DEBRECEN 2003 7 MOTION ENERGY DETECTION u u u Conventional statistical methods parameterized expressiveness emphasis complexity information content speed base frequency Principal component analysis of joint and angle flexions dI/dt Motion Energy = first derivate Time series Event detection/continuous 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 1 101 201 301 Motion Energy Detection 401 501 601 701 801 Frames DEBRECEN 2003 TIME DELAYED NEURAL NETWORKS u u u u Time series analysis complete exclusion of observer SNNS 4.1 network simulator - feed forward network (time delays) three layers receptive fields number of parameters translation invariant learning oft+Dt, t+2Dt,..., t+dDt curve forms Repeated random division of cases in 3 sets Set 1: learning with independent variable 50% Set 2: validation 10% Set 3: classification 40% x 1 x 2 x 3 x 4 t DEBRECEN 2003 8 WALKING AND GENDER IDENTIFICATION Walking signals gender identification and reproductive status 8m u u u u Method 104 males (26) 105 females (22) motion energy analysis (1.2 - 3 sec) Socio-demographic data, females cycle state TDNN classification for sex Correct classification 63% (Rating studies from literature: 64%) Conception probability with ,,classified as female" .44 p =.03 N spectators with ,,classified as male" .54 p =.02 1m Laufsteg 4m 4,5 m Kamera 6.5 6.0 5.5 5.0 4.5 4.0 3.5 10 15 20 25 30 35 40 45 50 55 60 Frames DEBRECEN 2003 WALKING AS A SIGNAL u u SEX GENDER DEBRECEN 2003 9 SOCIAL PERCEPTION: ZERO AQUAINTANCE u Perception at zero acquaintance is adequate Very thin slices 100 msec Exact personality description Coherent with self description Coherent with third party description u u First impression Prediction of outcome of interactions from first 10 seconds Prediction of personalities from 10 sec No consistent behaviour correlates found DEBRECEN 2003 Strangers u u Meet-Crosscultural Analysis Female deception avoidance and control Methods Repertoire differences sitting positions behavior frequencies gaze speech no model possible Motion Energy Detection strangers meet n=78 sociocognitions u u Movement and interest changes in movement quality small, slow and short events= female interest male's perception of conversation DEBRECEN 2003 10 Depression Depression and Movement heterogeneous disregulations of the biogenic amines lowered cerebral blood flow in regions responsible for motorics u Methods 25 women (23) structured interview with critical questions EWL, BDI (non-clinical measurements) motion energy detection (30 sec) TDNN classification ex emotion. stability self-assurance (good) mood -0.08 -0.02 -0.06 -0.07 -0.02 0.17 0.26 0.20 0.55* -0.21 em -0.06 -0.15 0.00 0.15 0.01 0.06 0.29 0.31 -0.21 sp 0.30 com -0.23 0.44* -0.30 0.56* -0.41* 0.51* -0.29 -0.19 -0.66* -0.34 -0.33 -0.45* 0.21 0.51* 0.30 0.30 0.27 0.17 Correlations of BDI-Scores Correlations of EWL-Scores Unique patterns Classification 45% happy introverted dissatisfied fearful depressed dreamy nervous 0.51* -0.10 DEBRECEN 2003 BDI 0.30 0.23 -0.59* 0.55* * = p .05, **= p .01, ***= p .001; two-tailed LAUGHTER:THE CHAMELEON EFFECT Randomly paired strangers of the opposite sex (n=48; age 18.5) 10 minutes interaction filmed through one way mirror Sociocognitions assessment of mutual interest Episodes of laughter only one episode per person no speech during laughter females n=28 males n=19 duration 0.72 - 3.8 sec u u u u DEBRECEN 2003 11 METHODS u u u u Digitisation 384 x 288 of episodes 25 frames/sec 22 khz 8-bit stereo Motion-Energy-Detection model free noise reduction difference between t1, t2 ... tn amount of movement in roi greyvalue difference in roi Power (amplitude*frequency) Multichannel analysis DEBRECEN 2003 ALLOMIMESIS u u Number of significant correlations with interest (F/M-INT p<.05) Female laughs SA Female movement and female power (M 1.4, Stddev 0.8): FINT n.s. MINT n.s. MA Female power and male movement (M 1.4, Stddev 0.7): FINT 0.44 MINT n.s. CT Female and male movement (M 1.5, Stddev 0.6): FINT n.s. MINT 0.48 u Male laughs SA : Male movement and male power (M 0.9, Stddev 0.9): FINT n.s. MINT n.s. CT Male power and female movement (M 1.3, Stddev 1.0): FINT 0.47 MINT n.s. Male and female movement (M 1.3, Stddev 0.6): FINT n.s. MINT n.s. DEBRECEN 2003 12 u u Sexual signaling and hidden ovulation Methods N= 123 females (age 23) reaction to a command non-pilltakers Mating status singles paired alone stimulus male/female saliva estrogen levels SHOWING OFF 3 2 1 high Column 1 0 -1 -2 0 3 50 100 150 2 middle Column 1 1 0 -1 -2 0 3 50 100 150 2 1 low Column 1 Conventional analysis 0 -1 no effect for duration ! Motion energy detection (1.8 - 2.9 sec) single females with male stimulus slower and more complex movements Reaction of stimulus person stimulus male reacts to ovulating paired female Neural network analysis classification high estrogene 100% 0 -2 50 100 150 Time/frames DEBRECEN 2003 PERSONALITY u u Personality as a signal for detection of behavioral tendencies innate traits: extroversion, openness, neuroticism, conscientiousness and agreeableness BAS and BIS theory of personality --> motor system Method 14 males (23) in interaction with females videotape big five Motion-Energy-Detection (30 sec after start) Correlations of movement parameters with the big five Openness Extroversion Conscientiousness Agreeableness Neurotiscism Size .02 .08 -.16 -.22 -.49* Sum -.54* .12 .41 .10 .17 Dur -.25 -.10 .02 -.08 -.12 Com -.12 .01 .10 -.09 -.14 Info -.12 .10 -.12 -.15 -.64* Num -.24 .01 .24 .20 .56* DEBRECEN 2003 13 RISK TAKING IN YOUNG MALES u u u u u u u Risk-seeking and risk-taking tendencies are premium information in interaction Risk-seeking is a stable personal trait Risk seeking depends from levels of B-MAO and sex-hormones Method N= 255 males (22.8) in discotheque and at the university Random chosen term from activity game at same level of difficulty Sensation Seeking Scale Form V Movement analysis Risk-seeking depends on mate seeking and alcohol consumption High-low risk seeking in non alcoholized subjects classification by TDNN 83% correct Alkohol consumption classification 69% correct DEBRECEN 2003 DANCING u u Dancing as self-presentation indicates reproductive potential motor control and coordination biomechanic linkages hormonal status Dancing as compatibility test matching behaviour assortative mating DEBRECEN 2003 14 Dance to Your Own Beat 31 males (25) 40 females (24) free dancing movements (5sec) demographic data hormonal status rating study for quantized displays (32 subjects) DEBRECEN 2003 ATTRACTIVE u MOVEMENTS u u u Interrater correlations .15 - .19 for erotic and attractive Cronbachs alpha .69 - .79 Sex classification by neural network 62% correct Non-pill taking females high emphasis, low complexity of upper body is attractive low speed and high complexity are classified as female Males high speed, high complexity, high expressiveness of upper body is attractive DEBRECEN 2003 15 SHYNESS u N=89 dancers, mean age 22 NEO-FFI PANAS RATING SHYNESS FEMALE .446* .524* MOTION COMPLEXITY .658* RATING SHYNESS MALE .354 u u Motion Energy Detection Rating study n=10 females and n=10 males 20 randomly selected dancers Adapted Neo FFI Panas Number of spectators .454* .442* u SELF RATING *.... significant: p < 0.05 DEBRECEN 2003 DYNAMIC THEORY OF COMMMUNICATION Analogue and parallel :The window to the soul Affective information is the prerequisite for cognitive functioning Affective signaling is not falsifiable Inference and the Shared Manifold Regulation of communication intention recognition mindreading prediction induction of affect - search spaces Evolution of self-deception New non linear communication models DEBRECEN 2003 16 INFERENTIAL COMMUNICATION MODEL INFORMATION INFORMATION CHANNEL SIGNAL SIGNAL SIGNAL RECEIVER RECEIVER RECEIVER MESSAGE CHANNEL CHANNEL MESSAGE MEANING INTERNAL STATES MOVEMENT RESONANCE PRIMARY AFFECTIVE REACTIONS REDUCTION OF DECISION SPACE INTENTIONS u Resonance and the primacy of affect MIRROR INTERNAL STATES DEBRECEN 2003 Ma n y thanks for your attention DEBRECEN 2003 17 ...
View Full Document

This note was uploaded on 05/12/2010 for the course PSYCHOLOGY clinical p taught by Professor Assistant during the Spring '10 term at École Normale Supérieure.

Ask a homework question - tutors are online