Workshop: ERP Testing

Transcription

1 Workshop: ERP Testing Dennis L. Molfese, Ph.D. University of Nebraska - Lincoln DOE NIH R01 HL NIH R01 DC NIH R41 HD47083 NIH R01 DA NASA SA NASA SA

2 Workshop Goals

3 Workshop Goals Concepts & Definitions

4 Workshop Goals Concepts & Definitions Common Practices

5 Workshop Goals Concepts & Definitions Common Practices Analysis Approaches

6 Workshop Goals Concepts & Definitions Common Practices Analysis Approaches Dealing with Artifacts

7 Workshop Goals Concepts & Definitions Common Practices Analysis Approaches Dealing with Artifacts Problem Solving

8

9 Workshop Daily Schedule ERP Theory, Methodology, Issues Electrode issues Artifacts Day 1 Equipment Videos - Unpack & Setup ERP System Electrode Net Application Net Station Operation

10 Workshop Daily Schedule Day 2 Preprocessing of ERP Data Data Management & Analyses Videos - Packing up ERP System

11 Ultimate Goal You Become An Independent Neuroscience Investigator who can: 1. Design & conduct independent studies. 2. Develop the Skills to run data analyses. 3. Draft and submit imaging manuscripts. 4. Develop grant applications. 5. Revolutionize your major field of study. 6

12 The Training Plan 1. Two-day ERP workshop 2. Experiment planning session(s) 3. Hands-on training on ERP equipment 4. Conducting YOUR experiment 5. Data Analysis Assistance 6. Manuscript Development Assistance 7. Grant Application Assistance 7

13 1.Overview of ERP Theory, Methodology & Issues. Why ERPs? Correlation with cognitive & physiological events Time resolution (ms) Spatial resolution Portability No age limits Useful with or without behavioral response Cost 8

14 General Methodology Principles

15 General Methodology Principles Same as in any research:

16 General Methodology Principles Same as in any research: Screen & control participant variables

17 General Methodology Principles Same as in any research: Screen & control participant variables Control stimulus & experimental factors

18 General Methodology Principles Same as in any research: Screen & control participant variables Control stimulus & experimental factors Data quality

19 General Methodology Principles Same as in any research: Screen & control participant variables Control stimulus & experimental factors Data quality Database

20 General Methodology Principles Same as in any research: Screen & control participant variables Control stimulus & experimental factors Data quality Database Data analyses

21 General Methodology Principles Same as in any research: Screen & control participant variables Control stimulus & experimental factors Data quality Database Data analyses Replication

22 ERPs History Definitions Electrodes Testing Issues Applications

23 Where we have come from s 11

24 Where we have come from s 12

25 Where we have come from... Oscilloscope Tracings & Photographs 13

26 Where we have come from s 14

27 Where are we now... 15

28 ERPs to CVC Words Below Average Readers Average Readers Above Average Readers

29 Event-Related Potentials

30 Event-Related Potentials ERP

31 Event-Related Potentials ERP Portion of Ongoing EEG

32 Event-Related Potentials ERP Portion of Ongoing EEG Time-Locked to Stimulus Onset

33 Event-Related Potentials ERP Portion of Ongoing EEG Time-Locked to Stimulus Onset Temporal Information

34 Event-Related Potentials ERP Portion of Ongoing EEG Time-Locked to Stimulus Onset Temporal Information Spatial Information

35 Event-Related Potentials ERP Portion of Ongoing EEG Time-Locked to Stimulus Onset Temporal Information Spatial Information Comparability across the lifespan

36 EEG Activity

37

38 Sampling (Digitizing) Rates

39 Sampling (Digitizing) Rates Brain Stem Evoked Response (BSER) 1-15 ms

40 Sampling (Digitizing) Rates Brain Stem Evoked Response (BSER) 1-15 ms Peak Duration ms

41 Sampling (Digitizing) Rates Brain Stem Evoked Response (BSER) 1-15 ms Peak Duration ms 5-7 peaks to resolve

42 Sampling (Digitizing) Rates Brain Stem Evoked Response (BSER) 1-15 ms Peak Duration ms 5-7 peaks to resolve Sample 1/5-1/10 ms

43 Sampling (Digitizing) Rates Brain Stem Evoked Response (BSER) 1-15 ms Peak Duration ms 5-7 peaks to resolve Sample 1/5-1/10 ms Middle Latency Response ms

44 Sampling (Digitizing) Rates Brain Stem Evoked Response (BSER) 1-15 ms Peak Duration ms 5-7 peaks to resolve Sample 1/5-1/10 ms Middle Latency Response ms Peak Duration 3-5 ms

45 Sampling (Digitizing) Rates Brain Stem Evoked Response (BSER) 1-15 ms Peak Duration ms 5-7 peaks to resolve Sample 1/5-1/10 ms Middle Latency Response ms Peak Duration 3-5 ms Sample 1/2-1 ms

46 Sampling (Digitizing) Rates Brain Stem Evoked Response (BSER) 1-15 ms Peak Duration ms 5-7 peaks to resolve Sample 1/5-1/10 ms Middle Latency Response ms Peak Duration 3-5 ms Sample 1/2-1 ms Cognitive components ms

47 Sampling (Digitizing) Rates Brain Stem Evoked Response (BSER) 1-15 ms Peak Duration ms 5-7 peaks to resolve Sample 1/5-1/10 ms Middle Latency Response ms Peak Duration 3-5 ms Sample 1/2-1 ms Cognitive components ms Peak Duration ms

48 Sampling (Digitizing) Rates Brain Stem Evoked Response (BSER) 1-15 ms Peak Duration ms 5-7 peaks to resolve Sample 1/5-1/10 ms Middle Latency Response ms Peak Duration 3-5 ms Sample 1/2-1 ms Cognitive components ms Peak Duration ms Sample 4-5 ms

49 Sampling (Digitizing) Rates Brain Stem Evoked Response (BSER) 1-15 ms Peak Duration ms 5-7 peaks to resolve Sample 1/5-1/10 ms Middle Latency Response ms Peak Duration 3-5 ms Sample 1/2-1 ms Cognitive components ms Peak Duration ms Sample 4-5 ms CNV - Contingent Negative Variation 2 S - 10 S

50 Sampling (Digitizing) Rates Brain Stem Evoked Response (BSER) 1-15 ms Peak Duration ms 5-7 peaks to resolve Sample 1/5-1/10 ms Middle Latency Response ms Peak Duration 3-5 ms Sample 1/2-1 ms Cognitive components ms Peak Duration ms Sample 4-5 ms CNV - Contingent Negative Variation 2 S - 10 S Peak Duration 5-10 S

51 ERPs - Extracellular

52 Event Related Potentials (ERPs)

53 Event Related Potentials (ERPs) Time-locked to an evoking or eliciting event or stimulus.

54 Event Related Potentials (ERPs) Time-locked to an evoking or eliciting event or stimulus. Sequence of serially activated "processes" (components) detected at the scalp (or some biological surface) as distinct positivenegative fluctuations.

55 Event Related Potentials (ERPs) Time-locked to an evoking or eliciting event or stimulus. Sequence of serially activated "processes" (components) detected at the scalp (or some biological surface) as distinct positivenegative fluctuations.

56 Event Related Potentials (ERPs) Time-locked to an evoking or eliciting event or stimulus. Sequence of serially activated "processes" (components) detected at the scalp (or some biological surface) as distinct positivenegative fluctuations. Measures:

57 Event Related Potentials (ERPs) Time-locked to an evoking or eliciting event or stimulus. Sequence of serially activated "processes" (components) detected at the scalp (or some biological surface) as distinct positivenegative fluctuations. Measures:! (1) peak latency from evoking stimulus onset (ms)

58 Event Related Potentials (ERPs) Time-locked to an evoking or eliciting event or stimulus. Sequence of serially activated "processes" (components) detected at the scalp (or some biological surface) as distinct positivenegative fluctuations. Measures:! (1) peak latency from evoking stimulus onset (ms)! (2) peak amplitude in microvolts µv

59 Event Related Potentials (ERPs) Time-locked to an evoking or eliciting event or stimulus. Sequence of serially activated "processes" (components) detected at the scalp (or some biological surface) as distinct positivenegative fluctuations. Measures:! (1) peak latency from evoking stimulus onset (ms)! (2) peak amplitude in microvolts µv (3) polarity (deflection from baseline to + or -)

60 Definitions for ERP displays x-axis horizontal abscicca time - ms y-axis vertical ordinate voltage amplitude - µv 23

61 ERP Nomenclature P2 P3 P4 P1 N1 N2 N3

62 ERP Nomenclature After Desmedt, 1974

63 Display Positive vs. Negative Up

64 Display Positive vs. Negative Up Arbitrary

65 Display Positive vs. Negative Up Arbitrary Tradition: 70% show Negative Up

66 Display Positive vs. Negative Up Arbitrary Tradition: 70% show Negative Up

67 Display Positive vs. Negative Up Arbitrary Tradition: 70% show Negative Up Creates some confusion in comparing work across studies

68 Display Positive vs. Negative Up Arbitrary Tradition: 70% show Negative Up Creates some confusion in comparing work across studies

69 Display Positive vs. Negative Up Arbitrary Tradition: 70% show Negative Up Creates some confusion in comparing work across studies Good to practice inverting waves to gain rapid visual recognition of peaks

70 Display Positive vs. Negative Up Positive Up Negative Up + -

71 Display Positive Up Hz 60 Hz Notch

72 Display Positive Down Hz 60 Hz Notch

73 Variations in ERPs Trial by Trial Variations in Amplitude & Latency Note amplitude 500 ms ALPHA 10 trials selected every 10 trials across 100 trials 30

74 Peak Latency Variations Produce Different Width Peaks Peak Amplitude Variations Produce Different Size Peaks More Latency Shift Less Latency Shift

75 ERP Temporal - Spatial Dynamics

76 ERP Temporal - Spatial Dynamics Assess effects that differ in:

77 ERP Temporal - Spatial Dynamics Assess effects that differ in:

78 ERP Temporal - Spatial Dynamics Assess effects that differ in: Time (ms)

79 ERP Temporal - Spatial Dynamics Assess effects that differ in: Time (ms) Scalp region distribution (2-D scalp surface space)

80 ERP Temporal - Spatial Dynamics Assess effects that differ in: Time (ms) Scalp region distribution (2-D scalp surface space) Dipole effects (Time and 3-D space)

81 Basic Measurements Amplitude Peak amplitude (maximum/minimum point) Mean peak amplitude (average # of points) Latency Peak latency (maximum/minimum point) Mean peak latency (average # of time points) Area (Under the Curve) Area for specific region 33

82 ERP Amplitude & Latency Measures

83 Basic Measurements Amplitude Peak amplitude (maximum/minimum point) Mean peak amplitude (average # of points) 35

84 Basic Measurements Latency Peak latency (maximum/minimum point) Mean peak latency (average # of time points) 36

85 Basic Measurements Area (Under the Curve) Area for specific region 50% area (midpoint) a b c c 50% 37

86 Peak & Latency Analysis

87 Peak & Latency Analysis Pros: Traditional approach Appears straight forward & logical Cons: - Peaks are not always clear - Developmental issues (changes in latency & amplitude) - Latency shift across scalp & subjects - Subjective judgments - Variations in criteria across journal reports - Very time consuming in training & execution - Replication problems within/across labs - Inter-rater reliability (typically not conducted/reported)

88 Peak & Latency Analysis

89 Sample Neonate Responses 41

90 ERPs & Averaging S/N = Signal-to-Noise-Ratio Individual (single trial) ERPs are VERY small - depends on age (e.g., ~.5 to 30µV) Amplifier/environmental noise are large - varies across amps & manufacturers and models - ~10 µv RMS (same size to 20x larger than single trial ERP) Thus, single trial ERPs can be OBSCURED by large electrical events, i.e., amplifier noise, environmental signals, artifacts TO SOLVE PROBLEM: Repeat same stimulus & average resultant single trial ERPs together to increase S/N ratio to improve ERP (signal) quality. 42

91 ERPs & Averaging Goff,

92 ERPs & Averaging ERP noise level varies with number of trials to create the average - square root law (conservative) - noise level = square root of the number of trials: 9 trials = 3 2 or 33.33% of signal could be noise 16 trials = 4 2 or 25.00% of signal could be noise 25 trials = 5 2 or 20.00% of signal could be noise 36 trials = 6 2 or 16.67% of signal could be noise 49 trials = 7 2 or 14.29% of signal could be noise 64 trials = 8 2 or 12.50% of signal could be noise 81 trials = 9 2 or 11.11% of signal could be noise 100 trials = 10 2 or 10.00% of signal could be noise 44

93 ERPs & Averaging Relation of Trials to Signal Noise 120 Number of Trials in Average # Trials per Average % of ERP that is Noise ERP Signal Improvement Trade-off between improving S/N and completing an experiment. 45

94 Average ERP obtained early during test differs from later in the same test period. First 25 Trails Trial #s combined to make average ERP Reference = linked mastoids Last 25 Trails 46

95 ERPs & Averaging The MORE trials presented, the better the S/N ratio. The MORE trials presented, the LONGER the test session. The LONGER the test session, the LESS LIKELY the infant/ child will complete session. The LONGER the test session, the LESS LIKELY later ERPs will resemble earlier trial ERPs. The REAL key to testing populations is to obtain the best S/N ratio without overtaxing the subject (e.g., infant, child, adult).

96 Another Way To Look At ERPs ba

97 Increasing Positive Voltage Yellow Red Purple Dark Blue Increasing Negative Voltage

98 Neonate ERP to Speech Syllable Yellow Red Purple Dark Blue

99 Adult ERP to Speech Syllable Yellow Red Purple Dark Blue

100 QUESTIONS??? 52

101 Dipoles

102 Dipoles a) Dipole used as description of ERP generation.

103 Dipoles a) Dipole used as description of ERP generation.

104 Dipoles a) Dipole used as description of ERP generation. b) Dipoles perpendicular to surface (since cortex folds, not necessarily perpendicular

105 Dipoles a) Dipole used as description of ERP generation. b) Dipoles perpendicular to surface (since cortex folds, not necessarily perpendicular to scalp surface).

106 Dipoles a) Dipole used as description of ERP generation. b) Dipoles perpendicular to surface (since cortex folds, not necessarily perpendicular to scalp surface).

107 Dipoles a) Dipole used as description of ERP generation. b) Dipoles perpendicular to surface (since cortex folds, not necessarily perpendicular to scalp surface). c) Reflects differences in soma and dendrite ion flow across cortical layers.

108 Dipoles

109 Dipoles d) Model activity.

110 Dipoles d) Model activity.

111 Dipoles d) Model activity. e) Activity at scalp not necessarily result of ion movements immediately below electrode.

112 Dipoles d) Model activity. e) Activity at scalp not necessarily result of ion movements immediately below electrode.

113 Dipoles d) Model activity. e) Activity at scalp not necessarily result of ion movements immediately below electrode. f) Caution: Dipoles generated in one hemisphere may generate higher shifts in other hemisphere.

114 Low GRTR Scores 1-dipole Model 200 ms Match Mismatch

115 High GRTR Scores 2-dipole Model 200 ms Match Mismatch

116 Are Dipoles Real? SENSE 3/1/01 5/2/01 Left Hand Right Hand Left Hand Right Hand

117 QUESTIONS??? Lantz, G., Grave de Peralta, R., Spinelli, L., Seeck, M., & Michel, C. M. (2003). Epileptic source localization with high density EEG: How many electrodes are needed? Clinical Neurophysiology, 114, Michel, C. M., Lantz, G., Spinelli, L., Grave de Peralta Menendez, R., Landis, T., & Seeck, M. (2004a). 128-channel EEG source imaging in epilepsy: Clinical yield and localization precision. Journal of Clinical Neurophysiology. Michel, C. M., Murray, M. M., Lantz, G., Gonzalez, S., Spinelli, L., & Grave de Peralta, R. (2004b). EEG source imaging. Clinical Neurophysiology, 115, Tucker, D. M., Luu, P., Frishkoff, G., Quiring, J. M., & Poulsen, K. (2003). Corticolimbic response to negative feedback in clinical depression. Journal of Abnormal Psychology, 112,

118 Digitizing Rate How fast to sample the ERP signal? Convention = 250 Hz (4 ms intervals) Ultimately dependent on signal characteristics 59

119 Nyquist's theorem: Analog waveform may be uniquely reconstructed, without error, from samples taken at equal time intervals. Sampling rate must be equal to, or greater than, twice the highest frequency component in the analog signal (3x works better). Example: 9 Hz wave sampled 9 times/sec = 1 Hz waveform

120 Nyquist - 9 Hz signal Sampled at 29 Hz Sampled at 14 Hz Yields 9 Hz Signal Yields 4.5 Hz Signal Alias - appear as more energy (higher amplitude) at lower frequency

121 Nyquist - Signal is sum of Sinusoidal Frequencies of 6.5, 10, 19 Hz Srinivasan, Tucker & Murias,

122 How Many Electrodes Should You Use? Depends on : Research Question. Availability of Equipment. Source Localization AND Scalp Distribution Studies ALWAYS require LARGE number of Electrodes adults = 256 infants = 128

123 Resolution of Scalp Signals Simulation of Infant & Child Scalp ERP Signals. Simulation of Adult Scalp ERP Signals. 64

124 If spatial sampling is too sparse, high spatial details will alias into low spatial frequencies, distorting topographic maps & source localization! 7 cm Srinivasan, Tucker & Murias, 1985

125 Nyquist The smallest topographic feature that can be resolved accurately by a 32-channel array is 7 cm in diameter - about the size of an ENTIRE lobe of the brain!!! 66

126 QUESTIONS??? 67

127 Impedance Before lab computers EEG quality depended on paper recorded signal. Noise from power lines (50 or 60 Hz) difficult to separate once introduced, Procedure involved abrading skin to achieve a scalp-electrode impedance < 5 kilo Ohms. Abrasion removes surface epidermal layer that has greater impedance than underlying tissue. 68

128 Impedance 69

129 Impedance Voltage (V) = Current(I) X Resistance(R) If Resistance increases, Current flow will decrease: V/R = I If Voltage increases, Current flow increases: V = C x R Current measured in Amps Voltage measured in Volts Resistance measured in Ohms Plumbing Analogy: Voltage ~ Water pressure (in a tank) Current ~ Flow Rate (from the tank) Resistance ~ pipe size (allowing water to escape from tank)

130 Impedance 71

131 Impedance High vs. Low Impedance Amplifiers 72

132 Impedance High vs. Low Impedance Amplifiers (elec) vs. (amp) High vs. High: 5x Low vs. Low: 5x Practice electric circuits: 73

133 Impedance Ferree, T., Luu, P., Russel, J. S., & Tucker, D. M. (2001). Scalp electrode impedance, infection risk, and EEG data quality. Clinical Neurophysiology, 112, Note: Watts = Amps x Volts

134 Electrode Paste vs. Collodian Adhesive paste EC2 vs. collodion for long-term scalp electrodes placement 40 patients 20: electrode placement on scalp with collodion - Group C (ollodian) 20: EC2 used - Group P(aste). impedance of electrodes measured after electrode placement (T1) and after 24 h of recording (T2), Application time calculated for all patients RESULTS: At each observation, group C showed mean values of electrode impedance significantly higher than group P Collodion: T1: 16.8 kohm; T2: 6.5 kohm EC2 Paste: T1: 2.4 kohm; T2: 4.0 kohm, p < 1 x 10(-5). 75

135 Electrode Paste vs. Collodian Time required to make montage and provide daily maintenance was significantly shorter in group P than in group C Collodion: 44.3 and 19.7 min EC2 Paste: 20.8 and 10.5 min, p < 1 x 10(-5). CONCLUSIONS: EC2 paste attaches scalp electrode in less time, with better recording quality as a result of lower electrode impedance values, than collodion. SIGNIFICANCE: EC2 paste can substitute for collodion in electrode placement for long-term video-eeg monitoring, with an optimal cost-benefit ratio in terms of recording performance, time consumption, & safety. 76

136 QUESTIONS??? 77

137 Filters

138 Filters ERPs (and EEG) are electrical signals that vary in their frequencies and amplitude.

139 Filters ERPs (and EEG) are electrical signals that vary in their frequencies and amplitude. Filter determines the way in which amplifier sensitivity changes as frequency is reduced.

140 Filters ERPs (and EEG) are electrical signals that vary in their frequencies and amplitude. Filter determines the way in which amplifier sensitivity changes as frequency is reduced. Frequency response - bandwidth of amplifier determined by its high & low frequency filters.

141 Filters

142 Filters D.C. Amplifier

143 Filters D.C. Amplifier Sensitivity does not change with decreasing frequency.

144 Filters D.C. Amplifier Sensitivity does not change with decreasing frequency.

145 Filters D.C. Amplifier Sensitivity does not change with decreasing frequency. Subject to very slow change of output level (drift).

146 Filters

147 Filters Low Pass Filter - attenuates HIGH frequency while saving or passing through the LOW frequencies (high frequency filter, high band pass filter)

148 Filters Low Pass Filter - attenuates HIGH frequency while saving or passing through the LOW frequencies (high frequency filter, high band pass filter)

149 Filters Low Pass Filter - attenuates HIGH frequency while saving or passing through the LOW frequencies (high frequency filter, high band pass filter) High Pass Filter - attenuates LOW frequency while saving or passing through the HIGH frequencies (low frequency filter, low band pass filter)

150 Amplifier Filter Settings

151 Amplifier Filter Settings Signals are reduced 50% already when frequency reaches setting depicted on most amplifiers.

152 Amplifier Filter Settings Signals are reduced 50% already when frequency reaches setting depicted on most amplifiers. Referred to as Half-Amplitudes

153 Amplifier Filter Settings Signals are reduced 50% already when frequency reaches setting depicted on most amplifiers. Referred to as Half-Amplitudes E.G., Setting on an amplifier of 2Hz and 30Hz means signal already reduced by 50% at filter boundaries.

154 Filters

155 Filters Filtering - sometimes represented as a Time Constant (TC)

156 Filters Filtering - sometimes represented as a Time Constant (TC)

157 Filters Filtering - sometimes represented as a Time Constant (TC) Describes how amplifier responds to a voltage change

158 Filters

159 Filters Voltage -> amplifier is changed.

160 Filters Voltage -> amplifier is changed. Amplifier output changed but gradually returns to baseline, producing a curve (exponential curve) that approaches its final value at a decreasing rate.

161 Filters Voltage -> amplifier is changed. Amplifier output changed but gradually returns to baseline, producing a curve (exponential curve) that approaches its final value at a decreasing rate. This curve has time constant (the time it takes for the AMPLITUDE to FALL to 37% of its INITIAL VALUE).

162 Filters As TIME CONSTANT (TC) increases, high pass filter frequency decreases (memorize ***)

163 Filters TIME CONSTANT = C Frequency = f Pi = /(2 x Pi x C) = f 0.159/C = f 0.159/0.3 = 0.5 Hz (cut off point of lowfrequency.) TC = 0.1, low frequency passed = 1.59 Hz TC = 0.5, low frequency passed = Hz TC = 1.0, low frequency passed = Hz

164 Filters: ERP amplitude and latency WILL change when applying different filters. )#$% )% "#(% "#'% "#&%./010234% )"56% )*56%,"56% "#$% "% )% *% +% ),% )-% $)% $*% $+%,,%,-% &)% &*% &+% *,% *-% ')% '*% '+% -,% --% ()% (*% (+% +,% +-% )")% )"*% )"+% )),% ))-% )$)% )$*% )$+% ),,% ),-% )&)% )&*% )&+% )*,% )*-% )')% )'*% )'+% )-,% )--% )()% )(*% )(+% )+,% )+-% $")% $"*% $"+% $),% $)-% $$)% $$*% $$+% $,,% $,-% $&)% $&*% $&+% 86

165 Filters: ERP amplitude and latency WILL change when applying different filters. High Pass Filter Low Pass Filter NOTE: Filters do NOT cut off the signal at filter settings!

166 Filters Hz

167 Filters Hz (60 Hz notch filter)

168 Filters Hz

169 Note: 60 Hz filter has no effect on ERP waveform if LOW PASS = 30 Hz 91

170 Filters As LOW PASS filter setting DECREASES, Peak Latencies will INCREASE (occur later) and slower frequencies will become more prominent in the ERP waveform as higher frequencies are filtered out (excluded). aka: Peak Latencies will occur later in the ERP waveform. 92

171 Filters Hz

172 Filters Hz

173 Filters Hz

174 Filters Hz

175 Filters As LOW PASS filter setting INCREASES, Peak Latencies will DECREASE and higher (faster) frequencies will become more prominent in the ERP waveform as higher frequencies are included (not filtered out). aka: Peak Latencies will occur EARLIER in the ERP waveform. 97

176 Filters Hz

177 Filters Hz Lower Low Pass gives Longer Latencies!!!

178 Filters Hz

179 Filters Hz (60 Hz notch filter)

180 Filters As HIGH PASS filter setting INCREASES, Peak Latencies will DECREASE and higher (faster) frequencies will become more prominent in the ERP waveform as lower frequencies are excluded (filtered out). Amplitudes will appear to decrease (get smaller). aka: Peak Latencies will occur EARLIER in the ERP waveform. aka: Peak Amplitudes will decrease in size. 102

181 Filters Hz (60 Hz notch filter)

182 Filters Hz (60 Hz notch filter)

183 Filters Hz (60 Hz notch filter)

184 Filters - signals change with filtering

185 Filters: Topography Hz (60hz)

186 Filters: Topography Hz

187 Filters: Topography.3-30 Hz

188 Filters: Topography.3-10 Hz

189 Filters: Topography.3-5 Hz

190 Filters: Topography Hz

191 Filters: Topography Hz

192 Filters: Topography Hz (60hz)

193 Filters: Topography Hz (60hz)

194 Filters: Topography Hz

195 Take Home Memory Different Filters produce different ERP waveforms Latency shifts Amplitude variations (positive & negative peaks) Slope changes Component structure impacted 117

196 CRITICAL When reading the literature ALWAYS pay strict attention to filter settings and gain settings used by investigators. DIFFERENT RESULTS with DIFFERENT FILTERS and GAIN (amplitude) settings. 118

197 QUESTION If 2 ERPs are collected but with different filter settings, which is the REAL data? Will the REAL ERP please stand up! 119

198 QUESTIONS??? 120

199 Adult Peak Components ERPs usually described in terms of Peaks (positive or negative) Latency (post stimulus onset) Duration (e.g., slow wave) Scalp topography (maximal peak location) Source (location within the brain) 121

200 Scalp Volume Conduction Current flow across the scalp Produces latency shifts from one part of scalp to another Also produces amplitude shifts across scalp Signals sum across the scalp large positive wave on scalp meeting large negative wave could sum to flat line! 122

201 EXPERIMENTAL DESIGN ISSUES: Types of Experiments

202 EXPERIMENTAL DESIGN ISSUES: Types of Experiments Common approaches

203 EXPERIMENTAL DESIGN ISSUES: Types of Experiments Common approaches Odd-ball tasks (P3 or P300)

204 EXPERIMENTAL DESIGN ISSUES: Types of Experiments Common approaches Odd-ball tasks (P3 or P300) Sentence completion (N400)

205 EXPERIMENTAL DESIGN ISSUES: Types of Experiments Common approaches Odd-ball tasks (P3 or P300) Sentence completion (N400) Mismatch negativity (MMN)

206 EXPERIMENTAL DESIGN ISSUES: Types of Experiments Common approaches Odd-ball tasks (P3 or P300) Sentence completion (N400) Mismatch negativity (MMN) Error-related negativity (ERN)

207 EXPERIMENTAL DESIGN ISSUES: Types of Experiments Common approaches Odd-ball tasks (P3 or P300) Sentence completion (N400) Mismatch negativity (MMN) Error-related negativity (ERN) Feedback-Related Negativity (FRN)

208 EXPERIMENTAL DESIGN ISSUES: Types of Experiments Common approaches Odd-ball tasks (P3 or P300) Sentence completion (N400) Mismatch negativity (MMN) Error-related negativity (ERN) Feedback-Related Negativity (FRN) Habituation

209 EXPERIMENTAL DESIGN ISSUES: Types of Experiments Common approaches Odd-ball tasks (P3 or P300) Sentence completion (N400) Mismatch negativity (MMN) Error-related negativity (ERN) Feedback-Related Negativity (FRN) Habituation Contingent Negative Variation (CNV)

210 EXPERIMENTAL DESIGN ISSUES: Types of Experiments Common approaches Odd-ball tasks (P3 or P300) Sentence completion (N400) Mismatch negativity (MMN) Error-related negativity (ERN) Feedback-Related Negativity (FRN) Habituation Contingent Negative Variation (CNV) Random order of presentation

211 Adult Peak Components: Some Descriptions P1 or P50 (auditory)

212 Adult Peak Components: Some Descriptions P1 or P50 (auditory) - Not always present - Occurs earlier over posterior than anterior scalp electrode sites - Larger amplitudes over frontal and/or central regions

213 Adult Peak Components P1 or P50 (auditory) - Distribution symmetrical over both hemispheres except for anterior temporal regions where larger amplitudes occur over left hemisphere; -Overall, peak amplitude and latency decrease with age to the point where the peak disappears (Coch, et al., 2002).

214 Adult Peak Components P1 or P50 (auditory) - Frequently associated with auditory inhibition in sensory gating paradigm where paired clicks presented at short ISIs. Amplitude of averaged ERP to second of paired clicks is typically reduced compared to averaged response to the first click. Magnitude of suppression commonly interpreted as neurophysiological index of sensory gating.

215 Adult Peak Components P1 or P50 (auditory) - Reduced suppression frequently reported for schizophrenic patients. - In some neuropsychiatric disorders (schizophrenia, mania), peak amplitude to paired stimuli approximately equal. - P1 latency clinically used to diagnose neurodegenerative diseases (multiple sclerosis, Parkinson s Disease).

216 Adult Peak Components P1 or P50 (auditory) - Buchwald et al. (1992) proposed that P50 response associated with ascending reticular activating system (RAS) and its post-synaptic thalamic targets. - Thoma et al. (2003) and Huotilainen (1998) independently localized sources of P50 in superior temporal gyrus using MEG approach. -Weisser et al (2001) co-registered auditory evoked potentials & magnetic fields (AEFs). The resulting equivalent dipole model for ERPs consisted of one source in auditory cortex of each hemisphere and a radially oriented medial frontal source.

217 Adult Peak Components P1 or P50 (visual) - Visual P1 differs from auditory P1 in terms of evoking stimulus, neurocognitive and neurophysiological mechanism, peak latency, scalp distribution, neural sources. - Visual P1 typically recorded in a checkerboardreversal task or similar light-flashes paradigms but can also be present for other visual stimuli (e.g., faces) & is largest over the occipital regions. - Negative peak may be present at same latency over frontal, central areas.

218 Adult Peak Components P1 or P50 (visual) - P1 amplitude generally varies with amount of attention in Posner s attention cueing paradigm & in spatial selective attention experiments. - P1 reflects suppression of noise because amplitude decreased for unattended locations but did not increase for attended stimuli. - P1 amplitude also increased when speed of response was emphasized, suggesting that P1 may also reflect level of arousal.

219 Adult Peak Components P1 or P50 (visual) - Sources identified using PET, BESA, and LORETA methods in ventral and lateral occipital regions (Clark, et al., 1996; Gomez, et al., 1994). - Suggests striate (Strik, et al., 1998) or extrastriate (posterior fusiform gyrus) origin (Heinze, et al., 1994). - Rossion, et al. (1999) in a face identification paradigm reported similar sources and sources in posteriorparietal regions, suggesting additional involvement of dorsal and ventral neural components.

220 N1 (N100) 133

221 N1 (N100) N1 typically occurs approximately 100 ms after stimulus onset. One of easiest components to identify regardless of specific analysis approach. Good convergence in findings based on analyses of PCA factor scores (Beauducel, et al., 2000), baseline to peak amplitude (Pekkonen, et al., 1995; Sandman & Patterson, 2000), and baseline to peak latency (Segalowitz & Barnes, 1993). 134

222 N1 (N100) N1 assumed to reflect selective attention to basic stimulus characteristics, initial selection for later pattern recognition, & intentional discrimination processing. Peak latency & amplitude depend on stimulus modality. Auditory stimuli elicit a larger N1 with shorter latency than visual stimuli (Hugdahl, 1995). 135

223 N1 or N100 (Auditory) Maximum amplitude over frontocentral areas (Vaughn & Ritter, 1970) or vertex (Picton, et al., 1974). Some studies differentiate into 3 different components with maximum amplitudes over temporal areas (latency 75 ms and 130 ms) & over vertex (latency 100 ms; McCallum & Curry, 80; Giard, et al., 94). Naatanen and Picton (1987) reviewed the 3 components of N1. Proposed that early temporal and vertex components reflect sensory and physical properties of the stimuli (e.g., intensity, location, timing in regards to other stimuli) while later temporal component are less specific and reflect transient arousal. 136

224 N1 or N100 (Auditory) NOTE, majority of studies treat N1 as single component occurring 100 ms after stimulus onset with maximum amplitude at the vertex electrode. N1 amplitude enhanced by increased attention to stimuli (Hillyard et al, 1973; Knight, et al., 1981; Ritter, et al., 1988; Mangun, 1995) increasing inter-stimulus interval (Hari, et al., 1982). 137

225 N1 or N100 (Auditory) N1 most likely generated by sources in primary auditory cortex in the temporal lobe (Vaughn & Ritter, 1970). MEG, BESA, and lesions studies consistently localize auditory N1 in superior temporal plane (e.g., Papanicolaou, et al., 1990; Scherg, et al., 1989; Knight, et al., 1988). However, several studies proposed additional sources in frontal lobe that could be activated from temporal lobe (e.g., Giard, et al., 1994). 138

226 N1 or N100 (VISUAL) Usually largest (maximum) over occipital region (Hopf, et al., 2002) or inferior temporal regions (Bokura, et al., 2001). Amplitude larger in discrimination tasks, but smaller if short ISIs. [** could disappear] N1 discrimination effect attributed to enhanced processing of attended location (Luck, 1995), not due to arousal because amplitudes are larger in tasks placing no emphasis on the speed of response. 139

227 N1 or N100 (VISUAL) Not affected by inhibition (no Go/No-Go response differences). Like auditory N1, visual N1 occurs at 100 ms over central midline sites & 165 ms over posterior sites. Anterior N1 = response preparation because eliminated if no motor response required. 140

228 N1 or N100 (VISUAL) Located visual N1 sources in inferior occipital lobe and occipito-temporal junction using a combination of techniques (MEG, ERP, and MRI), Hopf et al. (2002). However, Bokura et al., (2001) using the LORETA approach, identified additional sources of the visual N1 in the inferior temporal lobe. 141

229 QUESTIONS??? 142

230 P2 143

231 P2 Like N1 and P1, long considered obligatory cortical potential since it has low inter-individual variability and high replicability Identified in many different cognitive tasks including selective attention, stimulus change, feature detection processes, and short-term memory. P2 sensitive to stimulus physical parameters such as loudness. Participant differences such as reading ability also change P2 amplitude to auditory stimuli. 144

232 P2 (Auditory) P2 often occurs together with N1, yet peaks can be dissociated. P2 scalp distribution less localized than N1 & has its highest amplitude over central region. Temporal peak of P2 can occur over a broader latency range than the preceding peaks, ranging from ms. 145

233 P2 (Auditory) P2 can be double-peaked. Similar to N1, P2 has been consistently identified by analysis procedures: PCA factor scores (Beauducel, et al., 2000) Baseline to peak amplitude (Beauducel, et al., 2000; Sandman, & Patterson, 2000) Baseline to peak latency (Segalowitz & Barnes, 1993) 146

234 P2 (Auditory) Generators for auditory P2 thought centered mainly in primary & secondary auditory cortices. Both auditory N1 and P2 often represented by 2 dipoles: one in primary auditory cortex and one in secondary auditory cortex. Using BESA and LORETA to identify dipole locations for the N1/P2 component, Mulert et al. (2002) identified one in superior temporal region with a tangential orientation while second was located in temporal lobe with a radial orientation. These dipoles reflected primary and secondary cortices, respectively. 147

235 P2 (VISUAL) P2 amplitude increases with complexity of the stimuli. Topographic distribution of visually elicited P2 is characterized by a positive shift at the frontal sites around ms after stimulus onset and a large negativity, approximately 200 ms following stimulus onset at the occipital sites Using BESA dipole analysis, Talsma and Kok (2001) reported a symmetrical dipole pair localized in the inferior occipital (extrastriate) areas. Findings suggest that both topographic distribution and dipole position varied slightly when attending vs. not attending to visual images. 148

236 QUESTIONS??? 149

237 N2 150

238 N2 Influenced by features of the experiment, such as modality and stimuli presentation parameters. Shares some of its functional interpretation with mismatch negativity (MMN) because both indicate a detection of a deviation between a particular stimulus and the subject s expectation. However, unlike the MMN, the subject MUST pay attention to the stimuli. Ken Squires, et al. (1975) first reported this component. Ss viewed 2 stimuli. When the following image did NOT MATCH what was expected, a larger N2 occurred over frontal regions. 151

239 N2 N2 has multiple psychological interpretations including: orienting response (Loveless, 1983), stimulus discrimination (Satterfield, et al., 1990), target selection (Donchin, et al., 1978), reflecting task demands (Johnson, 1989; Duncan, et al., 1994). N2 has more inter-individual variation (Michalewski, et al., 1986; Pekkonen, et al. 1995). N2 is smaller in amplitude & shorter in latency for shorter ISIs (Polich & Bondurant, 1997). 152

240 N2 Topography N2 topographic distribution depends on sensory stimulus modality: Auditory elicit largest N2 amplitude at vertex. Scalp current density analysis indicate bilateral sources in supratemporal auditory cortex. Visual elicited highest N2 amplitude over preoccipital region. N2 to visual stimuli varied based on the stimuli type, such as written words, pictures of objects, or human faces. 153

241 N2 Sources Using intracranial electrodes placed directly on cortex, letter-strings of recognizable nouns produced N2 component at 4th occipital gyrus near occipitotemporal sulci. Pictures of complex objects, (cars, butterflies) resulted in N2 response over inferior lingual gyrus medially & middle occipital gyrus laterally. Effect not present for scrambled pictures. Face recognition tasks elicit N2 over fusiform gyrus & inferior temporal or occipital gyri just lateral to the occipito-temporal or inferior occipital sulci (see N170). Such differing distributions indicate N2 may reflect category-specific processing (Allison, et al., 1999). 154

242 N2 and Inhibition N2 associated with Go/No-Go paradigm, in which subject responds to some stimuli (Go trials), but inhibits response to another class of stimuli (No-Go trials). ERPs on No-Go trials are characterized by a large negative peak relative to the Go trials between 100 and 300 ms after stimulus onset (response inhibition??). N2 occurred both in relation to overt & covert responses, indicating that N2 Go/ No-Go effect not due only to motor responses. Instead, N2 present whenever responses must be interrupted. 155

243 N2 and Inhibition Amplitude and polarity of N2 inhibition response changes depending on the complexity of the task. In some instances, the Go/No-Go response has been reported as a positive peak, suggesting this pattern was due to large amplitude of the P300 in difficult tasks. N2 was larger when subjects have less time to respond. 156

244 N2 and Inhibition N2 for the visual & auditory task is especially strong over fronto-central electrodes when the Go response is withheld. Scalp distribution differs from Error Related Negativity (ERN) that occurs approximately 125 ms after an incorrect response. N2 response engages different processes than the error monitoring processes reflected in the ERN. 157

245 N2 and Inhibition Mathalon et al. (2003) using ERP and fmri identified activation of caudal and motor anterior cingulate cortices during both correctly and incorrectly inhibited responses. These sources differed from ERN responses that were related to caudal and rostral anterior cingulate cortices. Reinforces view the N2 reflects inhibitory responses distinct from error-related negativity. 158

246 QUESTIONS??? 159

247 Mismatch Negativity (MMN). 160

248 Mismatch Negativity (MMN). Naatanen et al. (1978) first described MMN wave as a negative deflection, latency = ms. Amplitude largest frontal & central electrode sites. MMN is elicited using an oddball paradigm where an occasional deviant stimulus is presented in a stream of more frequent standard stimuli. Test-retest reliability. Because MMN paradigms require no attention to the stimuli, widely used in developmental research. 161

249 Calculating the MMN Traditional: Subtract the averaged waveform of all standard stimuli FROM the averaged waveform of all deviant stimuli collected during the same test session. Alternative (2004): Present uninterrupted string of standard stimuli midway through experimental session to provide a alternative baseline for calculating the MMN. 162

250 Kraus, McGee, Carrell, Zecker, Nicol, & Koch, 1996

251 Mismatch Negativity (MMN) MMN evoked by any perceivable physical deviance from the standard stimulus (e.g., changes in tone duration, frequency, intensity, and ISI). Numerous theories Memory trace" - MMN elicited in response to violations of simple rules governing properties of information - violation of an automatically formed, short-term neural model or memory trace of physical or abstract environmental regularities Population of sensory afferent neuronal elements that respond to sound, and; ii) a separate population of memory neuronal elements that build a neural model of standard stimulation and respond more vigorously when the incoming stimulation violates that neural model "Fresh afferent" - sensory afferent neuronal elements that are tuned to properties of the standard stimulation respond less vigorously upon repeated stimulation. Thus when a deviant activates a distinct new population of neuronal elements that is tuned to the different properties of the deviant rather than the standard, these fresh afferents respond more vigorously. Sensory afferents are memory neurons. 164

252 Mismatch Negativity (MMN) Auditory MMN often used to test ability of subject to discriminate linguistic stimuli (e.g., speech sounds with different voice onset time or place of articulation. Data analyzed by subtracting average ERP elicited by standard stimuli from average ERPs for the deviants. This subtracted component generally displays an onset latency as short as 50 ms and a peak latency = ms (Naatanen, 1992). 165

253 Mismatch Negativity (MMN) Sources for auditory stimuli MEG: significant differences between dipoles produced by deviants differing in intensity, frequency and duration (Rosburg, 2003). Dipoles for frequency and duration deviants located significantly inferior in comparison to the source of intensity deviants and differed significantly from each other in the anterior-posterior direction. All dipoles located within temporal lobes. Leibenthal et al. (2003) recorded fmri and ERP data simultaneously to an MMN task. Main areas of increased BOLD signal in right superior temporal gyrus & right superior temporal plane. 166

254 Mismatch Negativity (MMN) Features influencing MMN Negative wave usually associated with MMN. Reports of positive wave around 200 ms corresponding to the MMN response (Leppanen, et al., 2002). The reason for this difference may be due to differences in filter settings. ** 167

255 Mismatch Negativity (MMN) Features influencing MMN Some reports indicate substantially reduced MMN response in subjects not attending to the stimuli Probability deviant stimuli influences effect. Must maintain balance between presenting enough deviant trials to obtain low-noise average responses, and not allowing the subject to habituate to the deviant, thus diminishing effect. Size of MMN response decreased (non linear), Time for habituation varies as function of stimulus complexity. 168

256 Mismatch Negativity (MMN) Visual MMN is found for visual stimuli (Tales, Newton, Troscianko & Butler, 1999). Source Localization techniques suggest involvement of primary visual cortex and adjacent areas (Gratton, 1997; Gratton, et al. 1998). 169

257 N170 Face Processing 170

258 N170 N170 ranges between 156 & 189 ms. Associated with visual processing of human faces. Topographic distribution for both familiar & unfamiliar faces largest over occipito-temporal regions. Amplitude significantly larger when viewing faces than other natural or human-made objects. 171

259 N170 Prosopagnosia Patients do not show an N170 response to faces. N170 not specific to human faces but expert object recognition (Tanaka & Curran, 2001) Intracranial recordings of EP & fmri point to fusiform gyrus as neuroanatomical substrate of N170. BUT source localization of N170 using BESA identified source in lateral occipitotemporal region outside fusiform gyrus. 172

260 QUESTIONS??? 173

261 P300 - Two Components P300a component associated with the automatic 'Orienting Reflex' P300b component associated with controlled processing (most studied) 174

262 P300 Odd-ball tasks S. Sutton, M. Braren, J. Zublin, and E. John, (1965) Evoked potential correlates of stimulus uncertainty Science, 150,

263 P300 Odd-ball tasks P300 amplitude increases to infrequent stimulus S. Sutton, M. Braren, J. Zublin, and E. John, (1965) Evoked potential correlates of stimulus uncertainty Science, 150,

264 P300 Odd-ball tasks P300 amplitude increases to infrequent stimulus Frequent 80% of trials, infrequent 20% S. Sutton, M. Braren, J. Zublin, and E. John, (1965) Evoked potential correlates of stimulus uncertainty Science, 150,

265 P300 Odd-ball tasks P300 amplitude increases to infrequent stimulus Frequent 80% of trials, infrequent 20% Requires attention & response to infrequent stimulus S. Sutton, M. Braren, J. Zublin, and E. John, (1965) Evoked potential correlates of stimulus uncertainty Science, 150,

266 P300 Odd-ball tasks P300 amplitude increases to infrequent stimulus Frequent 80% of trials, infrequent 20% Requires attention & response to infrequent stimulus Controls important S. Sutton, M. Braren, J. Zublin, and E. John, (1965) Evoked potential correlates of stimulus uncertainty Science, 150,

267 P300 Odd-ball tasks P300 amplitude increases to infrequent stimulus Frequent 80% of trials, infrequent 20% Requires attention & response to infrequent stimulus Controls important ERP averages based on same # trials for both frequent and infrequent stimuli S. Sutton, M. Braren, J. Zublin, and E. John, (1965) Evoked potential correlates of stimulus uncertainty Science, 150,

268 P3a 176

269 P3a P3a, frontal maximum scalp distribution. Slightly shorter latency for visual vs. auditory and somatosensory stimuli. Frontal P3a occurs when subject not required to actively respond to the targets (N. Squires, et al., 1975) or when novel stimulus is added to the standard 2-stimulus oddball paradigm. Frontal P3a assumed to reflect involuntary attention as well as inhibition. In Go/No-Go paradigms, P3a larger in amplitude in No-Go than Go conditions (maximal at parietal sites for Go). 177

270 P3a Neural substrate in medial parietal lobe (early: 317 ms) and in the left superior prefrontal cortex (late: 651 ms) for Go trials; Sources for No-Go trials (365 ms) originate in left lateral orbitofrontal cortex. P3a reduced by lesions to frontal cortex (Knight, 1991). 178

271 P3b or P

272 P3b or P300 Most extensively researched ERP component. Sutton et al., 1965: pronounced positivity occurring in response to unexpected stimulus approximately 300 ms after stimulus onset. Oddball most typical paradigm for eliciting P3b component, - a target stimulus presented infrequently among more common distracter stimuli. To get P3, subject must pay attention and respond to stimuli (unlike MMN) and the ratio of target to distracter stimuli must be low (fewer targets -> larger peak). 180

273 P3b or P300 AMPLITUDE affected by attention, stimulus probability, stimulus relevance, amount of processing resources available (e.g., single vs. dual tasks, quality of selection, and attention allocation. Interstimulus interval length affects AMPLITUDE independently of stimulus probability with shorter intervals resulting in larger P3b or P300. LATENCY varies with stimulus complexity, effectiveness of selection, and sustained attention. 181

274 P3b or P300 Visual P3 has larger & longer latency than auditory P3. P3 largest over parietal & midline regions. Auditory stimuli elicited shorter latency P3 over parietal regions, and longer latency over central sites. Functional interpretation of classic P3b diverse indicator of memory updating (Donchin & Coles, 1988) reflects a combination of processes that vary by task and situation, including more elaborate active stimulus discrimination and responses preparation. 182

275 P3b or P300 P3 latency assumed to reflect the duration of stimulus evaluation. P3 component attracted attention in clinical studies. Because P3 amplitude varies with the amount of attention paid to stimuli, this component widely studied in populations with attention deficits (e.g., ADHD) - interpreted to reflect information regarding various attentional functions. P3 latency reported related to cognitive abilities with shorter latencies associated with better performance 183

276 P3b or P300 Sources of P3 not clearly identified but some expected to be in medial temporal lobe, including hippocampal region related to memory (Donchin, 1981; Paller, McCarthy, et al, 1992), parahippocampal gyrus, amygdala, or thalamus (Katayama, et al., 1985). Lesion data suggest multiple generators, including temporoparietal junction (Knight et al, 1989). Tarkka et al., (1995) investigated possible sources and reported that combining different locations produced better model. MEG analyses located sources in floor of Sylvian fissure (superior temporal gyrus) and deeper sources in thalamushippocampus. 184

277 QUESTIONS??? 185

278 N400 He spread the warm butter with socks. 186

279 N400 Sentence Completion

280 N400 Sentence Completion N400 larger for unexpected, low probability endings.

281 N400 Sentence Completion N400 larger for unexpected, low probability endings. Fixed intervals between words

282 N400 Sentence Completion N400 larger for unexpected, low probability endings. Fixed intervals between words Words presented one at a time

283 N400 Sentence Completion N400 larger for unexpected, low probability endings. Fixed intervals between words Words presented one at a time Usual interval 1 S.

284 N400 Negative component approximately 400 ms after stimulus onset. Usually associated with semantic comprehension in both visual and auditory sentence comprehension tasks. First identified by Kutas and Hillyard (1979). Elicited by anomalies in American Sign Language. N400 did not occur when participants presented with anomalies in music (Besson, et al., 1994). 188

285 Kutas & Hillyard, 1980

286 N400 study with children The train runs on a track (CC) The train runs on a crack (CI) Child presses either red or green key to indicate if the sentence sounds ok or funny. 36 sentences for each condition. All sentences 6 words in length. Total data points digitized = 300

287 N year olds Word n=68 ms Incorrect Correct

288 N400 Paradigm Words of a sentence were visually presented one after another at fixed intervals in a serial manner. Last word of the sentence either congruous ( He took a sip from the water fountain ) or incongruous but syntactically appropriate ( He took a sip from the transmitter ) with rest of the sentence. Incongruous words elicited larger amplitude N400 response than congruous words for both auditory and visual stimuli. N400 amplitude correlated with degree of incongruency of final word to sentence (e.g., transmitter ) 192

289 N400 Kutas and Hillyard (1983): N400 effect only held true for semantic, but not syntactic deviations. Supposedly listeners use information from the wider discourse when interpreting appropriateness of particular word (van Berkum, et al., 2003). N400 also elicited in semantic word pairs (Rugg, 1985), semantic priming tasks (Bentin, et al., 1985; Ruz, et al., 2003) and matching semantic material to visual displays (Huddy, et al., 2003). 193

290 N400 (Modalities) For both visual and auditory displays, the N400 is larger for anomalous endings than expected endings over the parietal and temporal regions of the right hemisphere. But there are modality effects: N400 is earlier in the visual (475 ms.) than auditory (525 ms) modality but only over the temporal, anterior temporal and frontal sites (Holcomb, et al., 1992). Earliest peak in the visual modality is over parietal & temporal sites, while in the auditory modality it is over parietal & occipital sites (Holcomb, et al., 1992). 194

291 N400 (Asymmetries) Activation in left hemisphere occurrs earlier than activation in the right) in ONLY visual modality (Holcomb, et al., 1992). N400 not specific to written words, because spoken words (McCallum, et al., 1984; Holcomb, et al., 1992; Connolly & Phillips, 1994) & pictures (Nigam, et al., 1992) elicit N400. N400 response also elicited by incongruent solutions to mathematical multiplication problems (Niedeggen, et al., 1999). 195

292 N400 & Attention Still Unclear: Amount of attention necessary to produce N400, Cognitive processes involved (Osterhout & Holcomb, 1995). Holcomb (1988) reported N400 more robust when attention required but occurs when participants not attending to stimuli. 196

293 N400 & Attention However, Bentin et al. (1995) reported (dichotic listening task) that N400 was absent for material presented in unattended ear. Amount of effortful semantic processing required is unclear. Kutas and Hillyard (1993) reported N400 effect even in tasks not requiring semantic processing although Chwilla et al. (1995) found no N400 when attention not directed to meaning of stimuli. 197

294 N400 & Sources Likely multiple generators that are functionally (Nobre & McCarthy, 1994) and spatially (Halgren, et al., 1994; McCarthy, et al., 1995) segregated. Recent work points to parahippocampal anterior fusiform gyrus as generator (McCarthy et al, 1995). MEG studies pinpoint lateral temporal region as origin of N400 response (Simos, et al., 1997). Intracortical depth recordings using written words point to medial temporal structures near hippocampus & amygdala (Halgren, et al., 1994). 198

295 Late Positive Component (LPC) 199

296 Late Positive Component (LPC) Positive-going ERP component. Studies of explicit recognition memory. Largest over parietal scalp sites (mastoid reference). Begins approximately ms after stimulus onset. Duration = 200 ms ERP "old/new" effect. 200

297 Late Positive Component (LPC) S given list to learn. ERPs recorded to new list including old and new words. S to indicate old vs. new words. Typical larger LPC to old vs new words. Also done as continuous test - each trial S indicates if old vs. new item. 201

298 Late Positive Component (LPC) ERP & fmri indicate lateral parietal cortex, perhaps with medial temporal lobe and hippocampus. 202

299 QUESTIONS??? 203

300 ERN Error Related Negativity ERN reflects activity of a brain system that detects & corrects for errors.

301 ERN Paradigm Two ways to generate an ERN response: following an incorrect response during feedback of incorrect choice Hajcak, Holroyd, Moser, Simons, 2005; Holroyd, & Coles, 2002; Holroyd, Nieuwenhuis, Yeung, Cohen, 2003

302 ERN 206

303 ERN Paradigm During speeded response timing tasks, an incorrect response produces a negative peak ~ 100 ms Gehring, Goss, Coles, Meyer & Donchin, 1993 For reinforcement tasks, negativity around 250 ms indicates performance was incorrect Miltner, Baun & Coles, 1997 Negativity changes in amplitude for incorrect responses in high reward conditions or correct responses in low reward condition Holroyd, Nieuwenhuis, Yeung, & Cohen, 2003

304 ANALYSIS-Feedback Amplitude for ERN measured from baseline to peak between 160 ms to 240 ms following feedback Holroyd, Nieuwenhuis, Yeung & Cohen, 2003 Holroyd et al., (2003) used algorithm to identify amplitude of the greatest negativity in the peak starting at the slope of the first negativity through 325 ms Latency measures start at the maximum component amplitude

305 ANALYSIS - Incorrect Response Amplitude measured early: post incorrect response Luu et al., 200 Usually look at incongruent trials (i.e. Flanker task/go-no go task) Generate individual waveforms for error trials **Some groups used smoothing techniques with a nonphase-shifting single pass 17-point moving average (34 ms, approximately 3 db down at 15 Hz) --Santesso, Segalowitz & Schmidt, 2005 Filters set around 20 Hz offline Holroyd and Colleagues

306 ERN and Personality High Negative affect (high neuroticism) results in larger amplitude on initial tasks Tucker et al., 1999 ERN reflects certainty of loss (greater realization=greater ERN) Investment in task changes ERN Holroyd and Coles, 2002; Scheffers and Coles, year old children ranked on Junior Eysenck Personality Questionnaire show different ERN High psychoticism and low lie scores result in smaller ERN Similar to adults: see Dikman and Allen (2000) ERN affected by personality and concern of task performance Santesso, Segalowitz, & Schmidt,(2005)

307 Source Within the anterior cingulate cortex (ACC), mesencephalic dopamine neurons synapse on motor neurons and cause behavior to occur (e.g. pushing correct button) Basal ganglia mediated by feedback and stimulus input impact ERN generated by the electrical charge of neurons synapsing on the ACC. When events are worse than expected, decrease in dopamine activity disinhibits dendrites in ACC, resulting in a negative waveform, ERN

308 Source The Anterior Cingulate Cortex 212

309 Feedback-Related Negativity A negative deflection in the waveform approximately ms after the participant given negative feedback (Miltner, Braun, & Coles, 1997) Thought to originate in the anterior cingulate cortex (Ruchsow, Grothe, Spitzer, Kiefer, 2002).

310 Feedback-Related Negativity Generated without the individual having a choice in responding (Yeung, Holroyd, & Cohen, 2005) Generated without the individual responding (Donkers, Nieuwenhuis, & van Boxtel, 2005)

311 Feedback Negativity 250Hz

312 Feedback Negativity 250Hz

313 Scalp Topographies Win [Average: 19_7622f_rps.ref] Draw [Average: 19_7622f_rps.ref] Lose [Average: 19_7622f_rps.ref] 00:00: Win Draw Lose