Neural "ignition": enhanced activation linked to perceptual awareness in human ventral stream visual cortex

Abstract

Human recognition performance is characterized by abrupt changes in perceptual states. Understanding the neuronal dynamics underlying such transitions could provide important insights into mechanisms of recognition and perceptual awareness. Here we examined patients monitored for clinical purposes with multiple subdural electrodes. The patients participated in a backward masking experiment in which pictures of various object categories were presented briefly followed by a mask. We recorded ECoG from 445 electrodes placed in 11 patients. We found a striking increase in gamma power (30-70 Hz) and evoked responses specifically associated with successful recognition. The enhanced activation occurred 150-200 ms after stimulus onset and consistently outlasted the stimulus presentation. We propose that the gamma and evoked potential activations reflect a rapid increase in recurrent neuronal activity that plays a critical role in the emergence of a recognizable visual percept in conscious awareness.

Figure 1. Location of All ECoG Recording Sites (445 Electrodes in 11 Patients) Superimposed on a Cortical Reconstruction of One Subject from a Previous Study

Electrode locations as obtained from CT scans are indicated and color coded by amplitude of visually induced gamma power modulations in the first 250 ms after image onset in the Backward Masking experiment (yellow and orange for significant response, blue for none). Electrode positions are shown on a lateral view of the cortical hemispheres (top), an unfolded cortical surface (middle), and ventral view of the cortical hemispheres (bottom). The distorted shape of electrodes is due to the projection of the spherical shapes on the curved cortical surface. Occipito-temporal cortex, including fusiform gyrus and collateral sulcus, is shown enlarged. In these enlargements, electrode locations are shown relative to face-, man-made object-, and house-selective regions (colored contours), as delineated by fMRI localization; approximate retinotopic borders are indicated by dotted lines. “Low level” electrodes located in retinotopic cortex are marked by light-blue contours (see Experimental Procedures for further details).

Figure 2. Four Examples of Electrode Responses to Different Object Categories

Spectrograms showing induced responses to image presentation during the Object Category experiment. Colors indicate power increase (red) or decrease (blue) relative to prestimulus baseline. (A) Low-order visual electrode. (B–D) High-order, target-selective electrodes: man-made object-selective (B) and face-selective (C and D). Note that the low-order visual electrode does not show category selectivity and has short response latency, while the high-order electrodes show longer latency and clear category selectivity. The exact location of each electrode (number marked in top right corner of each panel, along with area) can be found in Figure 1 enlargements. IOS, inferior occipital sulcus; FG, fusiform gyrus; LO, lateral occipital. n = number of trials. Red horizontal bar indicates image presentation duration.

Figure 3. The Backward Masking Experiment

Figure depicts the images and time sequence used in the Backward Masking experiment. In each trial, a face or object image (including a fixation spot), the “target,” was presented for 16 ms, followed by a fixation-only blank of variable duration (SOA minus 16 ms). A scrambled image, the “mask,” was then presented for 250 ms. Subjects were allowed to respond verbally up to the beginning of the next trial. Overall trial duration was 3 s. The sequence was then repeated using a different target image. Note that only the SOA was parametrically manipulated in this experiment.

Figure 4. Neuronal Responses of Category-Selective Electrodes in the Backward Masking Experiment

(A) Gamma BLP (30–70 Hz band) response of four category-selective (three face-selective and one man-made object-selective) electrodes from three patients to 16 ms presentation of targets from the preferred category at the critical SOA (16, 33, 16, and 83 ms from left to right). Average responses when subjects succeeded or failed in recognizing the images are shown in red and blue, respectively. Mask-only responses are shown in green. SEM is given by shading. Time points showing significant recognition-related increase in gamma BLP (p < 0.01, corrected) are shown in yellow. (B) Control showing responses to the same exemplar composition but at longer SOA in which patients successfully recognized both sets. (C and D) Same analysis as (A) and (B) but for evoked responses. Numbers of trials are for (A) and (C). R, recognized; U, unrecognized; B, mask-only trials. Note the significant and long-lasting increase in gamma BLP (A) and the increase in the evoked (C) response amplitudes associated with successful recognition for all electrodes shown (see text). No such enhancement was seen when the same image compositions were compared during successful recognition of both sets (B and D).

Figure 5. Average Induced Response Spectrograms of Category-Selective Electrodes in the Backward Masking Experiment

(A–D) Spectrograms of four electrodes (same electrodes as in Figure 4) showing induced power changes in response to face presentation in the Backward Masking experiment. Correct recognition, recognition failure, and mask-only conditions are shown for each electrode. Time of face and mask presentation are indicated by red and stippled bars, respectively, below the top spectrograms. Electrode number and location are indicated in the top right corner. Color scale as in Figure 2. Note the sustained enhancement of gamma power (mainly above 50 Hz) as well as the transient low-frequency power rise associated with the face recognition condition. n = number of trials.

Figure 6. Average Evoked and Gamma Power Responses of All Target-Selective Face Electrodes

(A) Grand average of evoked target responses in the Backward Masking experiment around the critical SOA (including 16, 33, and 66 ms; see Experimental Procedures). Means are shown for recognition (red) versus no-recognition (blue) conditions; yellow region indicates time points in which the two responses differed significantly (p < 0.01, paired t test). The trial composition of both conditions is identical in terms of target stimulus identity, but not SOA. Error bars denote SEM across electrodes (n = 14). (B) Control data depicting mean evoked target responses at different SOAs (16, 33, and 66 ms; n = 215, 341, and 113 trials, respectively), for recognized trials only. (C and D) Same analysis as (A) and (B), respectively, but for gamma BLP. Note the significant enhancement in both evoked (A) and induced gamma (C) responses during correct recognition trials. This enhancement cannot be attributed to the changes in SOA since similar changes in SOA did not produce enhanced activation as long as the patients were able to correctly recognize the target images (B and D).