Support: NIH HD51912The Jerold B. Katz Foundation

The James S. McDonnell FoundationContact: zma2001@med.cornell.edu

B3CNS 2017

Methods

Schematic of a forward vs. time-reversed language paradigm. Forward and time-reversed English language excerpts were taken from the opening 2.5 minutes of Lewis Carroll’s Alice in Wonderland with female narration. These were presented with an interleaved rest of 30 seconds. Each language block, totaling approx. 6 minutes was presented a minimum of 2x’s per test session (healthy controls) and a minimum of 3-4x’s for patients.

RestForwardLanguage

ReversedLanguage

ALICE PARADIGM

1- 4 repeatsless EMG artifactparadigm run during arestricted time period(between 1-5 pm)

Healthy ControlsN = 15

1- 6 repeatsmore EMG artifactparadigm run multipletimes over the courseof an in-patient visit

Patient SubjectsN = 18 N = 14

patient subjects withanalyzable Alice data

N = 7patient responders via

TGT and FDR correctionin at least 1 channel/run

N = 9patient responders viaTwo Group Test (TGT)

in at least 1 channel/run

Background & Motivation• Language-based EEG paradigms can identify covert cognitive

processes in patients with disorders of consciousness (DOC)(Conte, SFN 2015; Markell, SFN 2014).

• Desynchronization in the alpha frequency band (8-12 Hz) is amarker of auditory attention (Weisz et al., 2011; Banerjee etal., 2011).

• Here we attempt to further characterize the modulation of theEEG in response to narrative language as well as thetopographic distribution of responses in both healthy controls(HCs) and patient subjects (PSs).

Sample Spectra - Healthy Controls

FWDBKWD

0 4 8 12 16 20 24 28 32 36 40

AF7F7

FC5T3

CP5T5

PO7

Fp1F3F1

FC1C3

CP1P3O1

FPzFzCz

CPzPz

POzOz

Fp2F4F2

FC2C4

CP2P4O2

AF8F8

FC6T4

CP6T6

PO8

Pow

er (d

B)

0 4 8 16 20 24 28 32 36 40−40

−30

−20

−10

0

10HC 1

12

POz

0 4 8 12 16 20 24 28 32 36 40

AF7F7

FC5T3

CP5T5

PO7

Fp1F3F1

FC1C3

CP1P3O1

FPzFzCz

CPzPz

POzOz

Fp2F4F2

FC2C4

CP2P4O2

AF8F8

FC6T4

CP6T6

PO8

0 4 16 20 24 28 32 36 40−40

−30

−20

−10

0

10HC 2

O2

128

Frequency (Hz) Frequency (Hz)

Pow

er (d

B)

PS 2: MCS-

Frequency (Hz)

Pow

er (d

B)

PS 1: VS

FWDBKWDBASELINE

Sample Spectra - Patient Subjects

LR

PS 3: MCS+ PS 4: E-MCS

LR

A. B.

LRLR

M, 25 yo; studied6 yrs. post-severetraumatic brain injury(MVA). In minimallyconscious state(CRS-R Tot. Score: 14).

F, 19 yo; studied 3 yrs. post-severetraumatic brain injury(MVA). In minimallyconscious state (CRS-R Tot. Score: 23).

F, 22 yo; studied 6.5 yrs. post-severetraumatic brain injury(MVA). In vegetative state, (CRS-R Tot. Score: 7).

M, 47 yo; studied 30 yrs. post-severetraumatic brain injury(MVA). In minimallyconscious state (CRS-R Tot. Score: 10).

18FDG-PET - global hypometabolism ofthe frontolateral, temporal, parietaland occipital cortices and basalganglia with moderate preservation ofthe midfrontal cortices and thenucleus accumbens.

18FDG-PET – moderate hypometabolismin: sensorimotor striatum, left prefrontal,right temporal, temporal-occipital andprecuneus. Almost normal metabolism:right striatum. Low metabolism: leftcentral thalamus

0 4 8 12 16 20 24 28 32 36 40−40

−30

−20

−10

0

10

20

POz

Pow

er (d

B)

0 4−40

−30

−20

−10

0

10

20

8 12 16 20 24 28 32 36 40

O1

0−40

−30

−20

−10

0

10

20

4 8 12 16 20 24 28 32 36 40

POz

0 4 8 12 16 20 24 28 32 36 40−40

−30

−20

−10

0

10

20

POz

Frequency (Hz)Power spectral responses to Fwd, Bkwd, and Baseline (eyes open, awake, no auditory stimulation) conditions from four patients withdistinct behavioral pro�les. Each panel illustrates data from a single repeat of the Alice paradigm. Baseline data in each PS wasconcatenated from several time periods over the course of one in-patient admission. Green stars indicate signi�cant di�erencesbetween the Fwd and Bkwd conditions. Signi�cant separations between Fwd and Bkwd conditions in the alpha frequency range aredemonstrated for MCS-, MCS+, and E-MCS patients; these di�erences remained signi�cant after FDR correction. No signi�cantdi�erences between the Fwd and Bkwd speech condition were observed in VS.

Continuous video EEG was recorded at a sampling rate of 250 Hz; bandpass-filtered (0, 40 Hz). We used the Natus XLTEK system for recordings with 37individual collodion-pasted Ag/Cl electrodes in an augmented 10-20 montage(Impedances < 5.0 kOhm).

Multi-taper (5 tapers) power spectral estimates were calculated using a HjorthLaplacian montage on artifact-free 3 sec EEG segments from single repeats of theAlice paradigm.

Power spectral significance differences between the conditions defined bycontiguous, Two Group Test (TGT) comparisons over the entire minimumfrequency resolution (2 Hz) of the multitaper estimate. All significant comparisonsunderwent False Discovery Rate (FDR) correction.

We recorded the EEG in 15 healthy controls (7M; age range 23-55) and 18 patientsubjects who sustained severe brain injuries. Participants listened to intact (FWDcondition) and time-reversed narratives (BKWD condition) within each recordingblock. The time-reversed BKWD condition is unintelligible but preserves theoverall power spectrum but not the phase information of the original speech.

18FDG-PET - shows hypometabolismin the left parieto-occipital and right lateral temporal region.

18FDG-PET image shown: no regions of severe hypometabolism. Not shown: hypometabolism in the left posterior occipital-temporal region.

ReferencesBanerjee et al. (2011). Oscillatory Alpha-band Mechanisms and the Deployment of Spatial Attention to Anticipated

Auditory and Visual Target Locations: Supramodal or Sensory-Speci�c Control Mechanisms? Journal of Neuroscience31(27): 9923-9932.

Conte, M.M., Fidali, B.C., Markell, H.M., Schi�, N.D. EEG Evidence of Auditory Working Memory and Selective Attention inDisorders of Consciousness, Program No. 405.18, 2015 Neuroscience Meeting Planner. Chicago, IL: Society forNeuroscience 2015. Online.

Markell, H.M., Mendels, L.F., Conte, M.M., Schi�, N.D. qEEG evidence for preservation of the auditory working memorybu�er in severely brain injured subjects, Program No. 703.02, 2014 Neuroscience Meeting Planner. Washington, DC:Society for Neuroscience 2014. Online.

Weisz et al. (2011). Alpha rhythms in audition: cognitive and clinical perspectives. Frontiers in Psychology 2(73): 1-15.

Conclusions Our results demonstrate the

preservation of languageprocessing in a subset of severelybrain-injured patients with limitedmotor output channels.

In both patient subjects andhealthy controls, the majority ofsigni�cant EEG responses occurredin centro-parietal and parietal-occipital regions consistent withactivation of cortical regionsinvolved in language processingand visual imagery.

Evidence of state �uctuations inindividual patients suggests theneed for repeated testing overmultiple testing blocks (see Curleyet al., this meeting).

Fluctuations in Wakefulness

This table provides evidence for diurnalstate �uctuations in the assessment ofpatients with DOCs. Highlighted cellscorrespond to the occurrence of threelanguage (2 passive, 1 active) paradigmsduring the testing day. For two patients,red circles indicate times during the daywhere we saw concordant positiveresponsiveness to these three paradigms.

Testing Block - Time of Day

Patient Subject 9am 12noon 3pm 6pmMCS- PS 2

Alice Day 2

Motor Imagery Day 2 Day 1

Personal Narratives Day 2 Day 1

MCS+ PS 3Alice Day 3

Motor Imagery Days 1 & 2

Personal Narratives Day 3 Day 3

E-MCS PS 4Alice Day 2

Motor Imagery Day 2

Personal Narratives Day 2 Day 2 Day 2

PS 5 (visit 1)Alice Day 1

Motor Imagery Day 2

Personal Narratives Day 1 Day 2

PS 5 (visit 2)Alice Day 1

Motor Imagery Day 2 Day 2 Day 2 Day 1

Personal Narratives Day 2 Day 1 Day 1

PS 6 (visit 1)Alice Day 2

Motor Imagery Day 3 Day 2 Day 3 Day 3 Day 3

Personal Narratives Day 1

0/15

3/15

6/15

0

Fp1

F3

FC1

C3

CP1

P3

O1

AF7

F7

FC5

T3

CP5

T5PO7

FPz

Fz

Cz

CPz

Pz

POz

Oz

FP2

F4

FC2

C4

CP2

P4

O2

AF8

F8

FC6

T4

CP6

T6PO8

F1 F2

Healthy Controls

CP6

T6

PO8

F1 F2

T4

CP6

T6

PO8

F1 F2

Fp1

F3

FC1

C3

CP1

P3

O1

AF7

F7

FC5

T3

CP5

T5

PO7

FPz

Fz

Cz

CPz

Pz

POz

Oz

FP2

F4

FC2

C4

CP2

P4

O2

AF8

F8

FC6

T4

Fp1

F3

FC1

C3

CP1

P3

O1

AF7

F7

FC5

T3

CP5

T5

PO7

FPz

Fz

Cz

CPz

Pz

POz

Oz

FP2

F4

FC2

C4

CP2

P4

O2

AF8

F8

FC6

T4

CP6

T6

PO8

F1 F2

Fp1

F3

FC1

C3

CP1

P3

O1

AF7

F7

FC5

T3

CP5

T5

PO7

FPz

Fz

Cz

CPz

Pz

POz

Oz

FP2

F4

FC2

C4

CP2

P4

O2

AF8

F8

FC6

Fp1

F3

FC1

C3

CP1

P3

O1

AF7

F7

FC5

T3

CP5

T5

PO7

FPz

Fz

Cz

CPz

Pz

POz

Oz

FP2

F4

FC2

C4

CP2

P4

O2

AF8

F8

FC6

T4

CP6

T6

PO8

F1 F2

0/9

3/9

5/9

Patient Subjects

Topographic Distributionof Signi�cant ResponsesN=15

N=9

L R

L R

Significant separations between Fwdand Bkwd conditions in alphafrequency band (8-12 Hz) power bychannel location in fifteen HCs (Two-Group Test; α < 0.05) (top left) andnine PSs (bottom left). Cold colorsrepresent channels where few or noHCs or PSs responded significantly,hot colors indicate channels wheresignificant responses occurredacross HCs and PSs. In the HCs, areaswith the highest density ofsignificant responses includeparieto-occipital, left centro-parietal, and fronto-central regions.In the PSs, activation was mostrobust in parieto-occipital and rightcentro-parietal areas. This biastowards significant responses on theright hemisphere in nine PSs reflectsa small sample statistic effect of leftposterior parietal-occipital injuries.

Spectral analysis of passive listening EEG paradigms reveals consistent patternsof activation in severely brain-injured patients

Zoe M. Adams 1, William H. Curley 1, Mary M. Conte 1, Nicholas D. Schi� 1,2,3

1Feil Family Brain and Mind Research Institute, Weill Cornell Medicine, NY; 2Department of Neurology, Weill Cornell Medicine, NY; 3The Rockefeller University, NY

Panels A and B illustrate sample HC powerspectral responses for a single repeat of theAlice paradigm in channels POz (left) and O2(right). Both HCs show signi�cant separationsbetween Fwd and Bkwd speech in the alphafrequency band (8-12 Hz) (as indicated by thegreen stars), with a Fwd > Bkwd directionality.Additionally, HC 1 illustrates a signi�cantdi�erence in the low beta frequency band (14-24 Hz). Below are shown the TGT analyses forthese conditions for all channels. Dashed redlines highlight channels POz and O2.Approximately half (Panel A: 49.94%; Panel B:43.34%) of TGT identi�ed values remainsigni�cant after FDR correction (not shown).