statistical aspects for the quantification of learning behaviour by sarah janssen ncs 2014, brugge 1...
TRANSCRIPT
1
Statistical aspects for the quantification of learning behaviour
By Sarah JanssenNCS 2014, Brugge
External supervisor: Dr. Tom Jacobs, Janssen Pharmaceuticals (J&J)Internal supervisor: Dr. Herbert Thijs, Uhasselt
2
Introduction
• A new animal behaviour model is setup to asses cognitive functioning in animals:– Animals are injected with PCP (also known as “Angel Dust”)– PCP has a degrading effect on learning behaviour– A good understanding of the effect of PCP on cognitive functioning is
important
• Optimizing the data analysis – That allows to quantify learning behaviour– That allows answering the research question in an unambiguous and
efficient way
3
The objective
• To study and quantify the dose effect of PCP on learning behavior
• To put it explicitly:– How does PCP affects learning behavior?– Which characteristics of learning behavior are sensitive to the dose
effect?– How to quantify the dose effect on these characteristics?– Which dose levels show a significant effect on learning behaviour?
4
Experimental setup
• Male wistar rats were trained to perform an action: choosing the correct image between two images
• Through reward mechanism• By the use of an operant
conditioning chamber• One training session ends after 48
trials or after 30 minutes maximally• Variable of interest: the proportion
of correctly executed trials within one training session
Figure Retrieved from www.campden-inst.com on 12/08/2012, URL: http://www.campden-inst.com/product_detail.asp?ItemID=1975&cat=2
5
Experimental setup
• Data available from two dose-response studies with PCP in identical conditions:– 96 animals – 5 dose levels: 0mg, 0.25mg, 0.5mg, 0.75mg, 1mg– Daily injection with PCP before every session– Sessions were performed daily during a period of 14 days
Dose level PCP: 0mg 0.25mg 0.5mg 0.75mg 1mg Total # of animalsTotal # of animals 24 12 24 12 24 96
6
Exploratory data analysis: individual profiles per dose level
0 2 4 6 8 10 12 14
0.0
0.4
0.8
Dose level: 0
days
prop
ortio
n
0 2 4 6 8 10 12 14
0.0
0.4
0.8
Dose level: 0.25
days
prop
ortio
n0 2 4 6 8 10 12 14
0.0
0.4
0.8
Dose level: 0.5
days
prop
ortio
n
0 2 4 6 8 10 12 14
0.0
0.4
0.8
Dose level: 0.75
dayspr
opor
tion
0 2 4 6 8 10 12 14
0.0
0.4
0.8
Dose level: 1
days
prop
ortio
n
7
• Variability between and within animals
• Profiles start around 0.5• Increase up to a level 0.9• Increase in a non-linear way• Less steep increase of the
profiles at higher dose levels
0 2 4 6 8 10 12 140
.00
.20
.40
.60
.81
.0
days
prop
ortio
n
Dose 0Dose 0.25Dose 0.5Dose 0.75Dose 1
Exploratory data analysis: average profiles per dose level
8
Part 1: Traditional Multivariate Anova model
9
The model
• Covariates: dose, time and dose*time• Residual errors are assumed to follow a multivariate normal
distribution• Pairwise comparisons of the 4 dose levels to the vehicle dose
at every time point• Without and with adjustment for multiple testing via
Bonferroni correction
ijkkjjkijk ε)(dose*timetimedoseμy
10
Results
11
Results
12
Conclusion
• Flexible model• Easy to understand and apply, also for non-statisticians• Inefficient way to analyze the data:
– Perform many test (59 comparisons)– Analyses becomes conservative when adjusting for
multiple testing• Does not answer the research question in a direct,
unambiguous way
13
Part 2: Non-linear mixed effects model
14
The model
• The response variable (proportion) is assumed to follow a beta distribution
• The average proportions (μij) are modeled as a Weibull learning curve (Gallistel et al, 2004):
2 4 6 8 10 12 14
0.0
0.2
0.4
0.6
0.8
1.0
days
prop
ortio
n
1
2
3 4
15
The model
• The Weibull distribution is characterized by a scale (L) and shape (S) parameter
• An intercept (I) and an asymptotic level (A) is added:
• To get a more meaningful interpretation for the scale parameter, L is reparameterized as T70:
• T70: time until proportion 0.7 was reached
(3) eIAI ]Ldayij
Sij )1(*)( )/[(
(4)
IAA
TL where
(3) eIAI
S
]Ldayij
Sij
/1
)/[(
847.0ln
70
)1(*)(
16
• This way, learning behavior is characterized by 4 parameters: – Intercept (I)– Asymptotic level (A)– Time to reach proportion 0.7 (T70)– Abruptness (S)
0 2 4 6 8 10 12 140
.00
.20
.40
.60
.81
.0
Panel A
days
prop
ortio
n
A=0.90
I=0.50
T70=5 days
S=3
The model
17
The model
• Dose effect is included in the model by allowing the parameters to change in function of dose level
• To take the heterogeneity between animals into account, random effects were included
(12) seA_slope*doA_A
(11) iI_I
(10) seS_slope*dosS_S
(9) e_slope*dosTt_TT
i*
i*
ii*
ii*
int
int
int
70int7070
18
Results
Parameter Estimate 95% CII_int 0.52 (0.50, 0.54)S_int 1.66 (1.42, 1.94)A_int 0.93 (0.92, 0.94)A_slope 0.81 (-0.13, 1.74)T70_int 3.9 (3.4, 4.6)T70_slope 1.11 (0.86, 1.36)
0 2 4 6 8 10 12 14
0.0
0.2
0.4
0.6
0.8
1.0
Model_L_2
days
prop
ortio
n
Dose 0Dose 0.25Dose 0.5Dose 0.75Dose 1
19
Results
Parameter Estimate 98.75% CI p-value
T700.25 / T700 1.14 (0.78, 1.67) 0.3883T700.50 / T700 1.50 (1.10, 2.06) 0.0014T700.75 / T700 1.61 (1.09, 2.36) 0.0022T701 / T700 3.21 (2.29, 4.49) <0.0001
20
Conclusion
• Weibull funtion was used to model the learning curves• Parameters have a biological interpretation• Direct, unambiguous answer to the research question:
– How does PCP affects learning behaviour? via T70– How strong is the dose effect 3 fold increase of T70 with a unit
increase of dose – Which does level show a statistical significant effect all, except dose
level 0.25
• Efficient way to analyze the data• Rather complex analysis
21
Thank you for your attention!
Thanks to:•Dr. Tom Jacobs, Janssen Pharmaceuticals (J&J)•Dr. Herbert Thijs, Uhasselt