Knowledge Vault 3/62 - G.TEC BCI & Neurotechnology Spring School 2024 - Day 5
Using multi-channel intracortical microstimulation to elicit tactile feedback
Charles Greenspon, University Chicago, 3rd place winner BCI Award (USA)
<Resume Image >

Concept Graph & Resume using Claude 3 Opus | Chat GPT4 | Llama 3:

graph LR classDef tactile fill:#f9d4d4, font-weight:bold, font-size:14px; classDef somatotopic fill:#d4f9d4, font-weight:bold, font-size:14px; classDef icms fill:#d4d4f9, font-weight:bold, font-size:14px; classDef multiElectrode fill:#f9f9d4, font-weight:bold, font-size:14px; classDef future fill:#f9d4f9, font-weight:bold, font-size:14px; A[Charles Greenspon] --> B[Touch critical for dexterous
bionic limbs. 1] B --> C[Skin highly innervated,
discriminates fine details. 2] B --> D[Force modulation requires
tactile feedback. 3] B --> E[Touch perceives texture,
softness through interaction. 4] A --> F[Somatotopic touch mapping
preserved in brain. 5] F --> G[Amputees' cortical hand
representation persists. 6] A --> H[ICMS produces sensations
referred to hand. 7] H --> I[Projected fields stable
over years. 8] H --> J[Mapping variability similar
within, across days. 9] I --> K[Projected fields align
with residual sensation. 10] H --> L[Individual ICMS electrodes
yield poor localization. 11] A --> M[Multi-electrode ICMS localizes
sensation to finger. 12] M --> N[Linear model predicts
multi-electrode projected field. 13] M --> O[Multi-electrode ICMS improves
finger stimulation accuracy. 14] M --> P[ICMS amplitude modulates
perceived intensity. 15] P --> Q[ICMS intensity equated
to finger indentation. 16] P --> R[ICMS amplitude discrimination
suggests intensity levels. 17] M --> S[Multi-electrode ICMS approximates
ecologically relevant forces. 18] A --> T[Biomimetic ICMS patterns
improve touch discriminability. 19] T --> U[Biomimetic multi-electrode ICMS
discriminates foam compliance. 20] A --> V[Multi-electrode ICMS produces
oriented edge percepts. 21] V --> W[Simultaneous multi-electrode ICMS
insufficient for letters. 22] A --> X[Sequential multi-electrode ICMS
produces motion percept. 23] X --> Y[Inter-electrode interval governs
perceived motion speed. 24] X --> Z[Sequential multi-electrode ICMS
improves spatial patterns. 25] A --> AA[ICMS safe, avoids
pain, damage. 26] A --> AB[Attention modulates near-threshold
ICMS detectability. 27] A --> AC[ICMS sometimes produces
'ghost' sensations. 28] A --> AD[ICMS 'hotspot' localization
time-consuming with many electrodes. 29] A --> AE[Multi-layer, high-density 3D
arrays may improve ICMS. 30] class B,C,D,E tactile; class F,G somatotopic; class H,I,J,K,L,AA,AB,AC,AD icms; class M,N,O,P,Q,R,S,T,U,V,W,X,Y,Z multiElectrode; class AE future;


1.- Tactile feedback is critical for dexterous use of bionic limbs, as vision alone is insufficient.

2.- Skin has high innervation density, especially in the fingers, allowing discrimination of fine spatial details.

3.- Force modulation when grasping objects requires tactile feedback. Vision alone leads to errors in applied force.

4.- Tactile feedback enables perceiving object properties like texture, softness/hardness through dynamic interaction with the object.

5.- Somatotopic mapping of touch is preserved throughout the nervous system from periphery to cortex.

6.- In amputees, cortical somatotopic hand representation persists, though some conflicting literature suggests remapping.

7.- Intracortical microstimulation (ICMS) of somatosensory cortex produces sensations referred to the hand (projected fields).

8.- Projected fields are stable over years in human participants, as measured by repeated mapping.

9.- Variability in mapping projected fields is similar within and across days, and is partly attributable to the mapping method.

10.- Projected fields align with receptive fields measured from residual sensation in some participants, supporting stability of cortical somatotopy.

11.- Individual ICMS electrodes yield poor localization of sensations in a rapid temporal order judgment task.

12.- Simultaneous stimulation through multiple electrodes with overlapping projected fields produces sensations localized to a single finger.

13.- A simple linear model can predict the projected field resulting from simultaneous stimulation of multiple electrodes.

14.- Multi-electrode stimulation dramatically improves participant accuracy in rapidly judging which finger was stimulated.

15.- ICMS pulse amplitude modulates perceived intensity of the tactile sensation.

16.- Perceived intensity from ICMS can be equated to indentation force applied to the finger.

17.- Amplitude discrimination thresholds for ICMS suggest 3-8 discriminable intensity levels per electrode.

18.- Multi-electrode stimulation increases the range of deliverable intensities to approximate ecologically relevant forces.

19.- Biomimetic patterning of ICMS trains to resemble neural responses to touch improves discriminability over linear amplitude modulation.

20.- Multi-electrode ICMS with biomimetic patterns enables discrimination of foam compliance, demonstrating the importance of patterning.

21.- Simultaneous activation of multiple electrodes with non-overlapping projected fields can produce percepts of oriented edges.

22.- Simultaneous multi-electrode stimulation is insufficient to reliably convey complex spatial patterns like letters.

23.- Sequential stimulation of multiple electrodes with short inter-pulse intervals produces a percept of motion across the skin.

24.- Temporal interval between sequentially activated electrodes governs perceived speed of motion across the skin.

25.- Rapidly sequential multi-electrode ICMS dramatically improves the ability to convey complex spatial patterns compared to simultaneous activation.

26.- ICMS of somatosensory cortex is safe and does not produce painful sensations. Parameters are chosen carefully to avoid damage.

27.- Attention can modulate detectability of near-threshold ICMS but has little effect on sensations at supra-threshold intensities.

28.- ICMS can sometimes produce "ghost" sensations, though these are difficult to distinguish from endogenous phantom sensations.

29.- Localizing the "hotspot" for ICMS is straightforward with a few electrodes but becomes time-consuming with large electrode counts.

30.- Efficacy of ICMS may be improved by targeting multiple cortical layers and expanding spatial coverage through high-density 3D arrays.

Knowledge Vault built byDavid Vivancos 2024