Dr. Gerald S. Hecht
Associate Professor of Psychology
College of Sciences
webmaster@psiwebsubr.org

PSYC 4035 - Drugs and Behavior Exam 2 Study Guide

CHAPTER 2: Research Design

- uses many subjects
- expensive
- fast
- group averages
- dependent variable only measured once

- uses fewer subjects
- slower
- dependent variable must be stable
- data from individuals
- no variation due to genetic and individual differences

The sum of the deviation scores divided by the total number of scores (observations)

Not very useful as a measure of dispersion because it is always equal to 0. Half of the scores will be below the mean (negative numbers) and the other half will be above the mean (positive numbers) so when you sum them they cancel each other out and equal 0. Remember a moment is a point of equilibrium between opposing forces—the mean itself is the measure of measure of central tendency in a normally distributed continuous data set—the first moment is 0 units of distance away from it—the mean itself is the point of equilibrium.

In order to get a useful measure of dispersion around the mean one can square the deviation scores prior to summing them. This will eliminate the negative signs from the scores that are below the mean thus yielding a quantity other than 0.
Importantly, scores that are farther away from the mean have a greater effect on the sum of squares than scores that are closer to the mean (e.g. a deviation score of 1 unit squared still equals 1 but a deviation score of 9 units squared equals 81. Therefore distributions that are very spread out will have much larger sums of squares than distributions that are less dispersed. Scores that are the farthest away from the mean will effect the sum of squares the most. At first, this seems like a useful measure of dispersion of a distribution—if all the scores cluster closely about the mean the sum of squares will be small. If the scores are widely distributed about the mean the sum of squares will be larger. It is tempting to use this as a measure of dispersion – small SS’s mean the distribution is less spread out and large SS’s mean it is very spread out. Unfortunately, however, the SS is also influenced by the number of scores in the distribution—the more of them there are the larger the SS will become. This is a source of contamination that precludes the use of SS as a pure measure of dispersion.
The second moment about the mean is also know as variance and is symbolized as

Although, we cannot use the SS as an uncontaminated measure of the dispersion of scores about the mean, we can divide the SS by the total number of scores and thereby calculate an average of the squared deviations. The important point here is that the variance averages the effects of large and small deviations from the mean and is therefore not affected by the number of observations in the data set. It is in fact restricted to your particular data set in the same way that the arithmetic mean as a measure of central tendency is. Therefore we can use the variance to characterize how much a randomly selected, typical score deviates from the mean… it is the first useful measure of dispersion in a normal distribution… it is our second moment.
Although the variance is a useful measure of dispersion, it is no longer based on the same units of measurement as our measure of central tendency (the arithmetic mean) because all of the deviation scores were squared before dividing them by the total number of scores. The variance is not interpretable in terms of our measurement scale. In order to correct the distortion in measurement resulting from squaring the deviation scores we can now calculate the square root of the variance. This measure of dispersion is known as the standard deviation
Armed with our formula, our measure of central tendency and first two moments about the mean we are able to calculate proportions of area under the normal distribution and thereby determine the probability (percentile ranking) of any given score on a measure of behavior (the dependent variable in Drugs and Behavior). In fact the proportion /probability of the area between the mean and + 1 standard deviation = .68 or 68 percent—- between + 2 standard deviations = .95 or 95 percent—and between + 3 standard deviations = .997 or 99.7 percent.

Skew refers to the symmetry of the distribution. A positive skew means that the arithmetic mean is larger than the median (from elementary stats chap 3) and a negative skew means that the arithmetic mean is smaller than the median. When calculating the magnitude of the third moment about the mean positive values = a positive skew and negative values = a negative skew. If the skew = 0 then the distribution is symmetrical and the mean = the median. Data sets with indices of skew that range between + 0.50 are considered to be within the limits of symmetry for calculations of area and percentile ranking with normal distributions.

The kurtosis is a descriptive measure that tells us whether the scores are bunched closely around the mean or spread out over a wide range with many scores in the “tails” of the distribution. Distributions that are spread out are described as being platykurtic, while those that cluster tightly about the mean are described as leptokurtic. The “perfect” normal distribution or classic “bell-shaped” curve is in between the platy and lepto kurtic extremes—it is known as mesokurtic and it occurs when = 3.
We have now reached the point where we can combine what we know about normal distributions to test various hypotheses regarding the behavioral effects (Dependent Variable) of Psychoactive Drugs (Independent Variable) in target populations. We will do this my calculating the effects of independent variables on dependent variables in samples and then use the principles of probability to make inferences about the populations from which our samples were drawn.
HYPOTHESIS: A statement concerning the relationship between independent and dependent variables. We will use several techniques from the field of formal logic to construct our hypothetical (i.e., concerning the hypothesis) arguments.
The methods of science used in hypothesis testing never prove anything (despite the cartoon version of science we get exposed to in movies, TV, etc.). Science uses the formal logical procedure of modus tollens to evaluate hypotheses. This is a procedure of falsification based on the fact that a single observation can lead to the conclusion that the premise or prior statement is incorrect. In psychological research the modus tollens is constructed through Null Hypothesis Significance Testing (NHST). In NHST two statistical hypotheses are created; the null hypothesis (H0) and the alternative hypothesis (H1). The null hypothesis is presented as the prior statement that is the “target” of our “falsification attack”. Logically, it is similar to the American criminal justice model—in which the condition of innocence is the a priori or null. Only if evidence accumulates during the trial that is inconsistent with that hypothesis being true, is the null rejected. In science, our null hypothesis takes the form of the argument that our Independent Variable (IV) has no effect on our Dependent Variable (DV). The null hypothesis specifies that the values of one or more population parameters (DV) do not differ as a function of exposure to an IV. The alternative hypothesis specifies that the population parameter under study (DV) is some value other than the one specified under the null hypothesis. If, at the end of our experiment, the probability of our DV values (scores—hint, hint) is extremely low if the null is really true, we reject the null hypothesis as false and tentatively accept the alternative hypothesis as being a more accurate reflection of reality.
Result X has occurred.
H0 is probably not true.
H0 :
H1:
The null hypothesis states there is no difference between the two means. If the null hypothesis is false the researcher should find that difference between the estimated population means is too large to be accounted for by chance (random) level probability alone. Therefore, the behavioral pharmacologist has reason to say the evidence does not support the null hypothesis and it should be rejected in favor of the alternative hypothesis being tentatively accepted as a more accurate description of the population parameters.
NHST decisions are based on probabilistic rather than definitive information. The first step therefore, is to determine a probability level where we are willing to reject the null hypothesis and tentatively accept the alternative hypothesis as being more consistent with the data. This point is known as the alpha level and the cutoff point as the significance level.
ALPHA ( ) LEVEL: The level of significance used by the experimenter to decide if the probability of obtaining certain results on a DV are so unlikely to occur by chance alone that they warrant rejecting the null hypothesis.
SIGNIFICANCE LEVEL: The probability value used to conclude that the null hypothesis is an incorrect statement. By convention, .05, .01 and .001 are frequently used significance levels in Psychological research. Failure to reject the null hypothesis does not mean that the null hypothesis is correct, it only means that we have not collected sufficient information to reject it as a false statement. We also have not proved that the alternative hypothesis is true. When we reject the null hypothesis, we are saying that the alternative hypothesis provides a more probable explanation for the data.
PROBABILITY (p) AND ALPHA ( ): If: p is less than or equal to alpha then: reject the null hypothesis in favor of the alternative hypothesis. If: p is greater than alpha then: fail to reject the null.
Sometimes you reject the null hypothesis and it was really true. You messed up (Remember those results could happen by chance 5 times out of 100 if your alpha was set at p=.05). At least you can know the name given to your mistake—a type I error (also known as an alpha error).

Other times you fail to reject the null hypothesis but the null hypothesis really is false—you messed up again—this time you have made a type II error (also known as a beta error).
STEP 1: State the null hypothesis. H0:

STEP 2: State the alternative hypothesis. H1:

STEP 3: Identify the appropriate statistical test. Student’s t-ratio for independent groups.

STEP 4: Determine the significance level. By convention

STEP 5: Identify the appropriate sampling distribution. Remember all of this is used to make inferences about the populations from which the samples were drawn.

STEP 6: Identify the critical region for rejection of H0. These values will come from what you know about probability and the normal curve, sampling distributions and your significance level.

STEP 7: Summarize and Analyze the data.
Pooled Variance Formula:  

t-Test Denominator Formula: 

t-Test Formula itself:
1. The Basic Scenario:
 
2. The reason for pooled variance considerations:
 
3. What you have actually determined:
 
Note that in the "real world" it would never be this simple -- remember our old friend known as the DRC (dose response curve) from the last unit. It would never be the case in a behavioral pharmacology experiment that a single dose of a drug would be compared with a placebo--you would have a range of doses tested to generate the DRC. Since the actual alpha level would double with each successive t-test (separately comparing each dose with the placebo control), your results would be contaminated by family wise alpha level inflation. You would need to use more sophisticated statistical techniques such as ANCOVA (using the placebo condition measures as covariates, etc.).
Placebo: inert substance given as though it were a drug

Placebo Effects:
- in humans drugs often do what people expect. Behavioral Pharmacology experiments must have a placebo control


Double-Blind Design - neither the participant nor the experimenter can know who
is getting the drug and who is getting the placebo.

1. New drug
2. Placebo
3. Established treatment
1X2 = tests hypothesis that New drug has significant effects.
2X3 = tests sensitivity of measurement methods to be used in null hypothesis tests.
1X3 = tests for advantages (i.e., potency, efficacy, TI, side-effect profile) over established treatment.

- includes case study method and correlational methods
- at best: can demonstrate correlation, not causality.
- unstructured introspection - "bad" (low reliability & validity). Example: "yo dude...I don't think this bud is as killer as that last bag...definitely way more headachy"
- structured introspection - "good" (can be very high in both reliability and validity). Example: Methods used in Psychophysics ("eye exam" introspection).
- two examples of "paper and pencil" instruments that are 1) examples of high quality structured introspection and 2) are highly sensitive to the effects of psychoactive drugs

1. POMS - Profile of Mood States
2. ARCI - Addiction Research Center Inventory

- Level of Neural Arousal
- typically a "crude" electrophysiological measure of neural activity (EEG) used to explore:

1. Sleep - stages

2. arousal and mood

3. arousal and activity
EEG reveals four categories of these brainwaves, ranging from the most activity to the least activity. When the brain is aroused and actively engaged in mental activities, it generates beta waves. These beta waves are of relatively low amplitude, and are the fastest of the four different brainwaves. The frequency of beta waves ranges from 15 to 40 cycles a second. Beta waves are characteristics of a strongly engaged mind. A person in active conversation would be in beta. A debater would be in high beta. A person making a speech, or a teacher, or a talk show host would all be in beta when they are engaged in their work.

The next brainwave category in order of frequency is alpha. Where beta represented arousal, alpha represents non-arousal. Alpha brainwaves are slower, and higher in amplitude. Their frequency ranges from 9 to 14 cycles per second. A person who has completed a task and sits down to rest is often in an alpha state. A person who takes time out to reflect or meditate is usually in an alpha state. A person who takes a break from a conference and walks in the garden is often in an alpha state.

The next state, theta brainwaves, are typically of even greater amplitude and slower frequency. This frequency range is normally between 5 and 8 cycles a second. A person who has taken time off from a task and begins to daydream is often in a theta brainwave state. A person who is driving on a freeway, and discovers that they can't recall the last five miles, is often in a theta state--induced by the process of freeway driving. The repetitious nature of that form of driving compared to a country road would differentiate a theta state and a beta state in order to perform the driving task safely.

The final brainwave state is delta. Here the brainwaves are of the greatest amplitude and slowest frequency. They typically center around a range of 1.5 to 4 cycles per second. They never go down to zero because that would mean that you were brain dead. But, deep dreamless sleep would take you down to the lowest frequency. Typically, 2 to 3 cycles a second.

It is a well known fact that humans dream in 90 minute cycles. When the delta brainwave frequencies increase into the frequency of theta/beta hybrid brainwaves, active dreaming takes place and often becomes more experiential to the person. Typically, when this occurs there is rapid eye movement, which is characteristic of active dreaming. This is called REM, and is a well known phenomenon.

These patterns of EEG activity used as dependent measures of neural arousal are, again very sensitive to the effects of psychoactive drugs (dare I say it again... the Independent Variable).

- Perception (Psychophysical Methods)
1. Method of Adjustment

    * The participant adjust the intensity of the stimulus until they can either just detect or just not detect a stimulus
    * Least accurate, but the fastest
    * Example – Person turns a knob until they can just sense tone or just not sense a tone
    * Example – using the contrast button or knob on your computer

2. Method of Limits

    * The participant is given discrete steps in order controlled by the experimenter
    * Example – eye test
    * Example, have 8 glasses of water, each with 2 cups of water and different levels of sugar. Glass 1 has no sugar, glass 2 has ¼ teaspoon (t) of sugar, glass 3 has ½ t of sugar, glass 4 has 1 t of sugar, etc. and glass 8 has 2 t of sugar.
    * Present glasses to individual, starting at different points going from until participant can just taste or not taste sugar. Y means can taste, N means can’t taste.

Trial 1Trial 2Trial 3Trial 4
Glass 1N
Glass 2NNN
Glass 3NNNN
Glass 4YYNY
Glass 5YYY
Glass 6YY
Glass 7Y
Glass 8Y

    * Why go in both directions (ascending and descending)? To prevent and cancel out the following errors
          o Error of habituation – individuals who keep responding the same way, work on principle that the stimulus is going to be the same as the last stimulus
                + want to be absolutely sure tasted sugar before responding yes
          o Error of anticipation – individuals who jump the gun and change their responses quickly, work on principle that the stimulus is going to be different from the last stimulus
                + as soon as possibility that might has tasted sugar, say yes
 
3. Method of Constant Stimuli

    * The stimuli are randomly presented a number of times
    * Slowest, but most accurate
    * Example, instead of giving above glasses of water in order, present them to the participant randomly.
    * Hearing test

Weber’s Law (just noticeable difference)
• Relates size of JND to size of the standard.
• Difference threshold gets larger as standard gets larger but ratio remains the same
•
– Delta R = change in physical magnitude needed for difference detection
– R = magnitude of standard stimulus
– k = Weber fraction
• Smaller the fraction, better can discriminate
 
Common Weber Fractions
• Light intensity 0.079, Sound Intensity 0.048, Lifted Weight 0.022, Line Length 0.029, Taste 0.083, Electric Shock 0.013
 
Fechner’s formula
• Assume all JND are psychologically equal (60-61 change is same as 300-305 change)
* The relationship fits a logarithmic function for all of the senses S = k log R
* where S=change in sensation (JND),
* R is the magnitude of the physical stimulus
* k is the constant derived from Weber’s law :
– Perception, constant, intensity
– Log – the power to which 10 must be raised to equal that number (log 10 = 1, log 100 = 2)

- Attention
1. Critical Flicker Fusion (CFF)
The CFF task asseses the integrative capacity of the central nervous system (CNS), and more specifically, the ability to discriminate discrete 'bits' of sensory information. Subjects are required to discriminate flicker from fusion, and vice versa, in a set of four light emitting diodes arranged in a one centimetre square. The diodes are held in foveal fixation at a distance of one metre. Individual thresholds are determined by the psychophysical method of limits on four ascending (flicker to fusion) and four descending (fusion to flicker) scales . The mean of these four ascending and descending presentations gives the threshold frequency in hertz. CFF has been shown to be sensitive to a variety of psychoactive compounds. A decrease in the CFF threshold is indicative of a reduction in the overall integrative activity of the CNS.

2. Choice Reaction Time (CRT)
The CRT task is used as an indicator of sensorimotor performance, assessing the ability to attend and respond to a critical stimulus . Subjects place the index finger of their preferred hand on a central starting button, and are instructed to extinguish one of six equidistant red lights, illuminated at random, by pressing the response button immediately in front of the light as quickly as possible. The mean of fifty consecutive presentations is recorded in milliseconds as a response measure of three components of reaction time: recognition, motor and total reaction time. Recognition reaction time (RRT) is the time it takes for the subject to notice the light, being the time between stimulus onset and the subject lifting their finger from the start button. Motor reaction time (MRT) indexes the movement component of this task, and is the time between the subject lifting their finger from the start button and touching the response button. The total reaction time (TRT) is the sum of RRT and MRT. CRT is sensitive to a variety of psychoactive agents.
- Vigilance
1. Continuous Tracking Test (CTT)
This interactive task of psychomotor function entails using a joystick control to keep a cursor in alignment with a moving target on a VDU screen. The movement of the target is a function of of an irregular sine wave. The response measure (RMS) is the mean difference between the centres of target and cursor in pixels, sampled 10 times per second, over the15 minute test. Lower scores are indicative of more accurate tracking. A peripheral awareness task (PRT) is included in which the subject responds to a stimulus presented in the periphery of vision, while simultaneously attending to the tracking test. The mean reaction time in milliseconds to these stimuli over the trial period is taken as the response measure for this component of the divided attention task.

- Motor Performance
1. Pursuit Rotor
Pursuit-Rotor Task The subject must keep the stylus in contact with
the metal disc that is moving in a circular pattern on a turntable,
which also is moving in a circular pattern. Although the task is
difficult, most people show significant improvement after a brief
period of training. Both initial performance and learning curves with repeated training
sensitive to various psychoactive drugs.2. Hand Steadiness

The subject’s task is to hold a metal-tipped stylus in 9 progressively smaller hole sizes without touching the sides. The effects on steadiness of right vs. left hands, resulting from administration of psychoactive drugs can be observed and accurately measured.

- Driving
1. Real Traffic
'nuff said.
2. Driving Simulator Tests Example: Brake Reaction Time (BRT)
BRT is used as a measure of cognitive and psychomotor performance including attentional efficiency. It is assessed in an automatic, dual control car, on a closed circuit. A red lamp mounted on the bonnet simulates the brake light of an imaginary vehicle in front of the test car. The lamp is illuminated at random intervals throughout the test. The subject is required to extinguish the lamp as quickly as possible, by depressing the brake pedal. The interval between the illumination of the red light and the initiation of braking is measured in milliseconds. A microcomputer is used to control the presentation of the stimuli and to record the results. Each test consists of 25 BRT trials. The mean latency of BRT following the illumination of red lamp is recorded. BRT is sensitive to pychotropic drug effects and provides results, which are consistent with laboratory measures of reaction times (CRT) and psychomotor performance.

1. SMA - Spontaneous Motor Activity.
The Animal Activity Monitor incorporates highly sensitive electronic capacitive sensors to detect changes in vibratory movements on the platform of the animal cage as a function of psychoactive drug ingestion/administration. The animal is placed inside an acrylic transparent cage that rests on the sensor platform. Testing can be conducted in complete darkness,or in bright sunlight. Ambient illumination is not a limiting factor as in most photocell activity monitors.It detects ambulatory movements as well as stereotypic activity like grooming, scratching, digging, etc. Vibrations caused by the animal activity produce proportional electrical signals. These are electronically processed to generate trigger pulses and drive a Digital Counter. Every count registered is accompanied by a flashing LED and beeper. The sensitivity of the animal activity monitor is adjustable as per user’s experimental requirements. An optional Timer is available to programme experiment  times from 1 to 99 minutes.

- Respondent Conditioning

1. Pavlovian conditioning example: Conditioned Place Preference (CPP)

The apparatus consists of two compartments that vary along several sensory dimensions, including the characteristics of the floor and the pattern or color of the walls. After each compartment has been repeatedly paired with drug or saline administration, the reinforcing or aversive effect of the drug is determined in a test session in which the animal has access to both compartments, and the amount of time spent in each is monitored with photocells or by visual observation.

- Operant Conditioning


1. schedules of reinforcement
  • Fixed Ratio Schedule (FR): A set number of correct responses must be made to obtain a reinforcer
  • Variable Ratio Schedule (VR): Varied number of correct responses must be made to get a reinforcer
  • Fixed Interval Schedule (FI): The first correct response made, after a certain amount of time has elapsed, is reinforced

2. Avoidance Operants:
Signal --> Response -->No shock
Signal --> No Response-->Shock

3. Escape Operants (Sidman Avoidance):
Shock-->Response--->no shock

4. Escape-Avoidance Operants:
Signal --> Response -->No shock
Signal --> No Response->Shock-->Response-->no shock

5. Punishment Operants:
phase 1: Response -->No shock + Reinforcement
phase 2: Response -->Shock + Reinforcement

6. Drug state discrimination:
Stimulus Properties of Drugs
- drugs can act as discriminative stimuli
- can be used to investigate new drugs
- can be used to investigate physiological mechanisms responsible for the
subjective effect of a drug.
- Summary of Differences Between Respondent (Pavlovian) and Operant Learning:
- Drug Development and Testing:

1. initial screening - nonhuman testing
ED50 and LD50

2. phase 1 - human testing
Toxicity and side effects

3. phase 2 - tested on patients

4. phase 3 - Clinical trials
3 groups design

5. phase 4 - extended data collection
often no compulsory reporting of side effects.